an improved mouldless manufacturing method for …€¦ · value analysis (pva). subsequently, the...

AN IMPROVED MOULDLESS MANUFACTURING METHOD FOR FOAM-CORE COMPOSITE SANDWICH STRUCTURES

By

Mario Mahendran

A thesis submitted to The Faculty of Graduate Studies and Research

in partial fulfillment of the degree requirements of Master of Applied Science

Ottawa-Carleton Institute for Mechanical and Aerospace Engineering

Department of Mechanical and Aerospace Engineering Carleton University

Ottawa, Ontario, Canada December 2010

Copyright © 2010 -M. Mahendran

1*1 Library and Archives Canada

Published Heritage Branch

395 Wellington Street OttawaONK1A0N4 Canada

Bibliotheque et Archives Canada

Direction du Patrimoine de I'edition

395, rue Wellington OttawaONK1A0N4 Canada

Your file Votre r$f6rence ISBN: 978-0-494-79537-8 Our file Notre reference ISBN: 978-0-494-79537-8

NOTICE:

The author has granted a nonexclusive license allowing Library and Archives Canada to reproduce, publish, archive, preserve, conserve, communicate to the public by telecommunication or on the Internet, loan, distribute and sell theses worldwide, for commercial or noncommercial purposes, in microform, paper, electronic and/or any other formats.

AVIS:

L'auteur a accorde une licence non exclusive permettant a la Bibliotheque et Archives Canada de reproduire, publier, archiver, sauvegarder, conserver, transmettre au public par telecommunication ou par I'lnternet, preter, distribuer et vendre des theses partout dans le monde, a des fins commerciales ou autres, sur support microforme, papier, electronique et/ou autres formats.

The author retains copyright ownership and moral rights in this thesis. Neither the thesis nor substantial extracts from it may be printed or otherwise reproduced without the author's permission.

L'auteur conserve la propriete du droit d'auteur et des droits moraux qui protege cette these. Ni la these ni des extraits substantiels de celle-ci ne doivent etre imprimes ou autrement reproduits sans son autorisation.

In compliance with the Canadian Privacy Act some supporting forms may have been removed from this thesis.

Conformement a la loi canadienne sur la protection de la vie privee, quelques formulaires secondaires ont ete enleves de cette these.

While these forms may be included in the document page count, their removal does not represent any loss of content from the thesis.

Bien que ces formulaires aient inclus dans la pagination, il n'y aura aucun contenu manquant.

14-1

Canada

ABSTRACT

An improved mouldless Liquid Composite Moulding (LCM) manufacturing method for foam-

core sandwich composite components was developed and utilized to manufacture a full scale

fuselage for the GeoSurv II Unmanned Aircraft System (UAS). Implemented as part of the low-

cost composite airframe research at Carleton University, this work intends to improve upon the

previously implemented mouldless Vacuum Assisted Resin Transfer Moulding (VARTM), using

effective design and manufacturing methods. Such methods will directly benefit small

aerospace companies, especially those dealing with general aviation aircraft and UASs.

In this work, numerous state-of-the-art LCM processes were reviewed. These identified Closed

Cavity Bag Moulding (CCBM) as a potential alternative to VARTM for mouldless manufacturing.

A series of experiments was carried out on CCBM process to evaluate its feasibility, including

investigations of various bag sealing and resin distribution strategies, followed by a Process

Value Analysis (PVA). Subsequently, the fuselage was redesigned, for improved

manufacturability. The new design was optimized using finite element analysis in Abaqus.

Finally, a full scale GeoSurv II fuselage was manufactured to demonstrate the viability of the

developed process. Results showed that this process is a viable option for manufacturing

complex foam-core composite components in small quantities. Improved part quality,

tolerances and weight reduction of 36% were achieved using the optimized design and

manufacturing method.

ABSTRACT ii

DEDICATION

Dedicated to my Mummy, Mary Rosaliya Mahendran and Dada, Kanesapillai

Julius Mahendran for their unconditional love and support.

DEDICATION i i i

ACKNOWLEDGMENTS

First of all, I would like to thank my supervisors, Prof. Paul V. Straznicky and Prof. Jeremy Laliberte for giving me this research opportunity and offering continuous guidance. As much as I appreciate their expertise, I must thank them for their exceptional patience throughout the project.

I would like to acknowledge Sander Geophysics Ltd. (SGL), NSERC Collaborative Research and Development (CRD) Grant and NRC-IAR for providing the financial resources for this research.

I also extend my appreciation to research officers, Dr. Chun Li, Drazen Djokic and Dr.Peter Krimbalis at the National Research Council-Institute for Aerospace Research (NRC-IAR) for providing consultation and support throughout my Master's program. I must also thank Tom Kay at NRC-IAR for his support in cutting the fibreglass rods.

My sincere gratitude goes to the technical staff of the Mechanical and Aerospace Engineering Department at Carleton University: to Alex Proctor for his support with machining of the foam cores, to Kevin Sangster for his guidance in the machine shop and to Steve Truttmann for his useful insights and support in the structures lab.

I also thank Aaron Miller (Composites Canada), John Biron (DIAB Inc.), Larry Audette (Prairie Technology Group Inc.) and Dave Kindt (Kindt-Collins Company LLC) for their interesting technical insights and support to this project. My appreciation also goes to Michel Reid, at Progress Plastics and Compounds Inc. for providing me with the opportunity to participate in the CCBM demonstration.

I am truly thankful to Frank Cappelli, Eileen Ruddek and Russ Elkin at 3A Composites/Baltek Inc., for their kind support and material donation.

Some of the experimental work in this thesis was completed with the assistance of several undergraduate and graduate students of the GeoSurv II UAS project. I am particularly thankful to Quinn Murphy (undergraduate summer student May-Aug. 2009) for his assistance with mechanical testing and documentation, Alexandre Adcock and Mauricio Buschinelli (GeoSurv II UAS project-2009/2010) for their support with the fuselage redesign work, Alan Lares (continuing graduate student) for his assistance with mechanical testing and fuselage manufacturing, Shashank Pant (continuing graduate student), who was always present when an extra hand was needed, Jeff Teutsch and Jerry Mac Pherson (GeoSurv II UAS project-2010/2011) for their assistance with characterizing the dimensions and quality of the fuselage. I am also thankful to Laurent Loisy, an exchange student from France, who provided valuable help with the final fuselage manufacturing.

My heartfelt gratitude goes to my parents and all members of my family for their love, care and support. Success would have been far from reach, without your role.

Special mention needs to be made for the wonderful human beings, who, through invaluable friendship, kept nourishing my life, both inside and outside of Carleton. Specifically:

• Jeeva, for all the wonderful moments we encountered at Carleton. You truly are a pillar of success in my academic career. Thanks for being around whenever needed, pointing me towards the right direction.

• Shashank and Fady, for the memorable coffee times and wing nights, where many exciting 'principles of life' were exchanged. My only regret is not getting to know you guys earlier in my career.

• Masih & Majed for being present 24/7 in the lab, exchanging ideas & sharing stories of life. • Dave, Henry and all other fellow graduate students in 2350 & Grad Lab • Mech & Aero Crew and others who stopped by the lab every now and then. • Sharmi & Co. for all the Quality Time. I won't thank you guys, but here, I'll say this one time 'you are my best gift'. • Not to be forgotten, I thank Nirth and Van for accompanying me in some labour intensive manufacturing tasks,

without a shadow of hesitation.

Last but not least, big thanks to everyone who periodically questioned the status of my thesis. Bluffing those questions was a definite inspiration, when intruded by other intellectual aspects of life.

ACKNOWLEDGEMENTS IV

CONTENTS

ABSTRACT ii

DEDICATION iii

ACKNOWLEDGMENTS iv

CONTENTS v

LIST OF TABLES ix

LIST OF FIGURES x

CHAPTER 1. INTRODUCTION 1

1.1. Background 1

1.2. Overview of the Low-Cost Composite Airframe Research 2

1.3. Overview of the GeoSurv II UAS 4

1.4. Thesis Objectives and Organization 6

1.5. Contributions 8

CHAPTER2. OVERVIEW OF THE GEOSURV II FUSELAGE 10

2.1. Design of the GeoSurv II fuselage 10

2.2. Manufacturing of the Fuselage Prototype 12

2.3. Post Process Assessment of the GeoSurv II Fuselage 13

2.4. Design and Manufacturing Objectives for the New Fuselage 15

CHAPTER 3. REVIEW OF LCM PROCESSES 17

3.1. LCM Processes 17

3.1.1. Resin Transfer Moulding (RTM) 20

3.1.2. Structural Reaction Injection Moulding (SRIM) 21

3.1.3. Vacuum Assisted Resin Transfer Moulding (VARTM) 22

3.1.4. Seemann Composite Resin Infusion Moulding Process (SCRIMP™) 26

3.1.5. Vacuum Assisted Process (VAP®) 26

3.1.6. Fast Remotely Actuated Channelling (FASTRAC) 27

3.1.7. Controlled Atmospheric Pressure Resin Infusion (CAPRI) 28

3.1.8. Double Bag VARTM 29

3.1.9. Advanced VARTM (A-VARTM) 30

3.1.10. Single Line Injection (SLI) 31

3.1.11. High Performance VARTM (Hyper-VARTM™/ Hyper-RTM™) 32

3.1.12. Vibration Assisted Liquid Composite Moulding 33

CONTENTS v

3.1.13. Closed Cavity Bag Moulding (CCBM) 33

3.1.14. Co-Injection Resin Transfer Moulding (CIRTM) 34

3.1.15. Euro Composites® Honeycomb Liquid Moulding (EC-HLM) 35

3.1.16. VacFlo® Process 37

3.1.17. Light RTM 37

3.1.18. Resin Infusion between Double Flexible Tooling (RIDFT) 38

3.1.19. Flexible Injection 40

3.1.20. Resin Film Infusion (RFI) 41

3.1.21. Semi-Preg Infusion 42

3.2. Suitability LCM Processes for Mouldless Manufacturing 43

CHAPTER 4. CCBM PROCESS DEVELOPMENT 46

4.1. Mouldless CCBM manufacturing considerations 46

4.1.1. Sealing Mechanisms for Mouldless CCBM 46

4.1.2. Resin Inlets and Outlets 47

4.1.3. Resin Distribution in CCBM 48

4.2. CCBM Manufacturing Trials 48

4.2.1. CCBM Manufacturing Trial # 1 49

4.2.2. CCBM Manufacturing Trial #2 51

4.2.3. CCBM Manufacturing Trial # 3 52

4.3. Value Analysis of the CCBM process 53

4.3.1. Introduction to Value Analysis 54

4.3.2. CCBM Process Value Analysis 55

4.3.3. PVA Results 56

4.3.4. PVA Conclusions 57

CHAPTER 5. FUSELAGE MATERIAL SELECTION 58

5.1. Sandwich Theory 58

5.2. Core Material for GeoSurv II Fuselage 60

5.2.1. Structural Foam Core Materials 61

5.2.2. Balsa Wood Cores 65

5.2.3. Other Core Materials 66

5.2.4. Core Selection for GeoSurv II Fuselage 67

5.3. Selection of Matrix and Reinforcement Materials 69

5.4. Material Selection for Rigid Inserts 70

CONTENTS vT

CHAPTER 6. FUSELAGE REDESIGN 72

6.1. Redesign objectives 72

6.2. Design Changes to the GeoSurv II Fuselage 73

CHAPTER 7. DESIGN OPTIMIZATION 80

7.1. GeoSurv II Fuselage FEA 80

7.1.1. FE Model Construction 80

7.1.2. Material Properties 82

7.1.3. Part Meshing Considerations 86

7.1.4. FE Model Assembly and Constraints 89

7.1.5. Analysis Steps, Loads and Boundary Conditions 90

7.1.6. Mesh Independence of the Results 93

7.1.7. FEA Simulations 95

7.1.8. FEA Results 97

7.2. Experimental Verification of the FEA results 102

7.2.1. Test Matrix, Specimen Manufacturing and Test Procedure 103

7.2.2. FEA Simulations 106

7.2.3. Results 108

CHAPTER 8. FUSELAGE MANUFACTURING 113

8.1. Sample Section Manufacturing 113

8.2. CCBM Experiments and Fuselage Manufacturing Model 117

8.3. Fuselage Manufacturing 121

CHAPTER 9. MANUFACTURING RESULTS 128

9.1. Surface Finish, Weight and Tolerances 128

9.2. Process Viability 131

CHAPTER 10. CONCLUSIONS 133

10.1. Conclusions 133

10.2. Recommendations for Future Work 134

REFERENCES 137

APPENDICES 141

Appendix A: Manufacturing Supplies 141

Appendix B: CCBM Bag Manufacturing Procedure 142

Appendix C: Process Value Analysis 145

Appendix D: Core Materials and Inserts 148

CONTENTS vii

Appendix E: FEA Results and Weight Estimates 157

Appendix F: Microscopic Image Analysis 158

Appendix G: Fuselage Profiling 160

Appendix H: Fibre Volume Fraction Calculation 168

CONTENTS viii

LIST OF TABLES

Table 2.1: Airframe requirements for the GeoSurv II UAS [5] 15

Table 2.2: Design and manufacturing objectives for the new fuselage 15

Table 3.1: Potential advantages of disadvantages of LCM processes [6-10] 18

Table 3.2: New materials and technologies developed to improve VARTM process 25

Table 3.3: Advantages and Disadvantages of CCBM process [24] 44

Table 4.1: PVA matrix: processes for mouldless CCBM/VARTM 55

Table 5.1: Manufacturing requirements for the core material 61

Table 5.2: Structural foam cores suitable for mouldless VARTM manufacturing 63

Table 5.3: Normalized mechanical properties of the most structural foam cores 68

Table 5.4: Carbon fibre fabric specifications 69

Table 6.1: GeoSurv II fuselage redesign: goals and limitations 72

Table 7.1: Properties of Airex C PVC foam 83

Table 7.2: Properties of the current fuselage materials 84

Table 7.3: Derivation of lamina properties for the new fuselage 85

Table 7.4: Properties of pins and bushings 86

Table 7.5: Parametric Study Results 94

Table 7.6: FEA verification test matrix 103

Table 7.7: Bearing test failure loads 108

Table 7.8: Bearing test results: FEA predictions and Experiments 109

Table 9.1: Comparison of major dimensions: fuselage design vs. current and new fuselages.. 130

Table 9.2: Comparison of the actual weight and predicted weight of the new fuselage 131

LIST OF TABLES ix

LIST OF FIGURES

Figure 1.1: Current configuration and specifications of GeoSurv II UAS (2009/2010) [4] 5

Figure 1.2: GeoSurv II Prototype Assembly (February 2010) 5

Figure 2.1: Main sections of the GeoSurv II fuselage [3] 10

Figure 2.2: Cross-section of the fuselage structural H frame [3] 11

Figure 2.3: Sequence of the main steps employed in mouldless VARTM 12

Figure 2.4: Fuselage main frame manufactured by mouldless VARTM 14

Figure 2.5: Fuselage prepared for assembly 14

Figure 2.6: Work-flow diagram of the new fuselage development 16

Figure 3.1: A family of state of the art LCM processes 20

Figure 3.2: Schematic of the Basic RTM setup 21

Figure 3.3: SSRIM process [8] 22

Figure 3.4: Schematic of basic VARTM process 23

Figure 3.5: Schematic of the EADS VAP® before (top) and after (bottom) infusion [10] 26

Figure 3.6: Schematic of the FASTRAC process [18] 27

Figure 3.7: Schematic of the CAPRI process [19] 28

Figure 3.8: Schematic of double bag VARTM 29

Figure 3.9: Outline of NCW fabric [20] 30

Figure 3.10: Sequence of steps in A-VARTM process [20] 31

Figure 3.11: Illustration of SLI process [21] 31

Figure 3.12: Pressure distribution during and after resin injection in SLI process [21] 32

Figure 3.13: Typical CCBM procedure 33

Figure 3.14: Schematic of the CIRTM method [25] 35

Figure 3.15: Details of the part during layup [26] 36

Figure 3.16: Schematic of the EC-HLM process [26] 36

Figure 3.17: Schematic of the RIDFT process [10] 39

Figure 3.18: Industrial RIDFT machine 10 ft x 15 ft x 4 ft [29] 39

Figure 3.19: Flexible Injection Process [30] 41

Figure 3.20: Schematic of the RFI process [31] 42

Figure 3.21: A part bagged with disposable vacuum bag [32] 45

LIST OF FIGURES x

:igure 4.1: Common reusable seal configuration for CCBM 47 :igure 4.2: Components of Arctek CCBM system [35] 49 :igure 4.3: CCBM bag manufacturing trial #1: tool setup 50 :igure 4.4: CCBM bag manufacturing: trial #1 results 50 :igure 4.5: CCBM bag manufacturing trial #2: tool setup 51 :igure 4.6: CCBM bag manufacturing: trial #2 results 52 :igure 4.7: Sample CCBM bag section with resin distribution channels 52 :igure 4.8: Flow profiles of CIB method (a) and disposable distribution media (b) 53 :igure 4.9: Cost of the fuselage for increasing part count at labour rate of 20$/hr 56 :igure 4.10: Cost of the fuselage for increasing part count at labour rate of 40$/hr 57 :igure 5.1: Sandwich beam subjected to three point bend 58 :igure 5.2: Sandwich beam in bending 59 :igure 5.3: Balsa wood core-end grain configuration [43] 65 :igure 5.4: Comparison of the selected foam materials at 4 lbs/ft3 density 68 :igure 5.5: Comparison of inserts for sandwich assembly 71 :igure 6.1: Fuselage redesign work-flow diagram 73 :igure 6.2: Current and the new fuselage wall design (units: in.) 74 :igure 6.3: Current and the new bolted sandwich assembly 75 :igure 6.4: New landing gear attachment plate 76 :igure 6.5: Current and the new landing gear configurations 76 :igure 6.6: Fuselage wall straight section extension 77 :igure 6.7: Design modification at the fairings 77

igure 6.8: Increased core thickness at the locations of the fasteners 78

igure 6.9: Core extension to mount the nosecone 78

igure 6.10: Reinforcement for flight avionics rack 78

igure 6.11: Current and the Modified fuselage concept models 79

igure 7.1: Parts modelled for the fuselage FEA 81

igure 7.2: Partitions created on the fuselage skin for meshing 88

igure 7.3: Mesh refinement near the pin holes 88

igure 7.4: Fuselage FE model assembly 89

igure 7.5: Tie constraints established between the pins and the fuselage structure 90

LIST OF FIGURES xi

:igure 7.6: Boundary Conditions 91 :igure 7.7: Wing lift and engine loads 91 :igure 7.8: Main landing gear loads 92 :igure 7.9: Mission avionics and nose landing gear loads 92 :igure 7.10: Substructure model used for parametric study 93 :igure 7.11: Mesh refinement: parametric study level 1 to level 12 94 :igure 7.12: Von Mises stress convergence (5%) 95 :igure 7.13: Fuselage FE model 96 :igure 7.14: Optimized foam core for the new fuselage 97 :igure 7.15: Optimized skin layup for the new fuselage 98 :igure 7.16: Von Mises (left) and in-plane shear (right) stresses (psi) in the skin under flight oads 100 :igure 7.17: Von Mises stresses (psi) in the skin during the landing step 101 :igure 7.18: Shear stresses (psi) in the skin under landing loads 101 :igure 7.19: Loading modes chosen for experiments 102 :igure 7.20: Geometry of the sandwich specimen 104 :igure 7.21: Bearing test setup in the load frame 105 :igure 7.22: Bending test setup in the load frame 106 :igure 7.23: Abaqus model showing the loads and boundary condition of the specimens 107 :igure 7.24: Abaqus FE model of the test specimens 107 :igure 7.25: Bearing test: force-displacement data 109 :igure 7.26: Comparison of the failure loads in bearing test 110 :igure 7.27: Close-up of the bearing failure mode 110 :igure 7.28: Bending test FEA prediction 112

igure 7.29: Bending test: force-displacement data 112

igure 8.1: Geometry of the test article 115

igure 8.2: Important features of mouldless CCBM setup 116

igure 8.3: Manufactured component 116

igure 8.4: Bondline comparison of VARTM and CCBM manufactured sandwich coupons 117

igure 8.5: CCBM experiment setup 118

igure 8.6: Conceptual CCBM Manufacturing Model 120

LIST OF FIGURES xii

Figure 8.7: Mouldless CCBM Process 121

Figure 8.8: Machining of the foam parts on the CNC router table 122

Figure 8.9: Foam parts required for fuselage manufacturing 123

Figure 8.10: FRP inserts for fuselage 123

Figure 8.11: Bonding of foam parts in the fixture 124

Figure 8.12: Features included in the foam parts to facilitate precise assembly 124

Figure 8.13: Mouldless CCBM setup 126

Figure 8.14: Resin starved regions observed during the infusion 127

Figure 9.1: Fuselage Manufactured by mouldless CCBM 128

Figure 9.2: Revised PVA cost estimates based on actual labour required for fuselage manufacturing (labour rate $40/hr) 132

LIST OF FIGURES xiii

CHAPTER 1. INTRODUCTION

This chapter provides an overview of the ongoing low-cost composite airframe research at

Carleton University. The structural demonstrator used in this research, the GeoSurv II

Unmanned Aircraft System (UAS) is also discussed. The thesis objectives are outlined, followed

by the procedure utilized to accomplish the objectives and the thesis contributions.

1.1. Background

Sandwich construction has been used in primary and secondary aircraft structures for many

years. It consists of thin face-sheets or skins adhesively bonded to both surfaces of a relatively

thick, low density core material. This leads to increased strength and stiffness for little added

weight. A properly designed sandwich construction also offers many other advantages such as,

thermal insulation, impact resistance and noise attenuation [1,2].

For many decades, aluminum and Nomex® honeycomb have been the two most commonly

used core materials in aerospace applications; they offer excellent specific strength-to-weight

and stiffness-to-weight ratios. However, their open and anisotropic cell structure leads to

several problems such as core crushing during shaping, curing or when subjected to lateral

forces. Honeycomb based sandwich components also exhibit extensive moisture ingress and

consequent corrosion, which lead to component service restrictions as well as increased life

cycle costs. Additionally, processing of honeycomb sandwich components typically requires

expensive autoclaves and the use of pre-impregnated composite fabrics (prepregs) [2].

Various foam and balsa core technologies have successfully addressed some of these issues at

much lower costs, with their closed cells and chemically resistant constructions. The


development of out-of-autoclave Liquid Composite Moulding (LCM) processes further enhances

the cost savings, making sandwich construction an economical solution for many structural

applications [3,4].

Developing low cost manufacturing methods without sacrificing part quality, repeatability and

performance, is always in the best interest of aerospace companies, specifically those dealing

with small aircraft and UASs for civilian applications. Costs for small aerospace companies can

be significantly reduced with the use of low cost manufacturing methods. This was the main

driving factor for the inception of low-cost composite airframe research at Carleton University.

1.2. Overview of the Low-Cost Composite Airframe Research

Optimizing the design and manufacturing methods to reduce cost has been an ongoing area of

research at Carleton University for the past four years. The research addresses one major

barrier to widespread use of composite materials in aircraft structures: high material and

manufacturing costs. Conventional prepreg layup and autoclave curing methods produce

composite parts with optimum fibre volume fraction, low void content and excellent surface

finish. However, the prepregs and autoclave infrastructure come with high initial and recurring

costs, which are beyond the means of small companies. Hence, the aerospace industry is

constantly searching for low cost, out-of-autoclave manufacturing methods capable of

producing high quality components. Such manufacturing methods will directly benefit small

companies, such as those that deal with general aviation aircraft and UAS [3,4].

The early stages of this research were funded by the Ontario Government, through the Ontario

Centre of Excellence (OCE) Interact program and Sander Geophysics Ltd. (SGL). Subsequent


financial support was also provided by SGL, along with an NSERC Collaborative Research and

Development (CRD) grant and support from the National Research Council - Institute of

Aerospace Research (NRC-IAR).

The low-cost composite airframe research uses the GeoSurv II UAS as a technology

demonstrator, to develop and implement the processes that are capable of producing

structural components of varying complexity. GeoSurv II is an excellent demonstrator, due to its

relatively small size and lower reliability requirements compared to a manned aircraft.

During the academic years 2006-2008, low cost processing of carbon-epoxy composite

components was investigated by Maley [3] and Vacuum Assisted Resin Transfer Moulding

(VARTM) was identified to be a suitable process for manufacturing most of the GeoSurv II UAS

components. Maley also suggested that manufacturing of sandwich structure components in

small quantities could be achieved economically through a single step "mouldless" infusion.

This process uses the core material as the mould to fabricate composite sandwich structures,

eliminating the material and labour costs associated with mould preparation. Hence, this

process can be very beneficial to aerospace companies for producing small aircraft and UASs in

low production quantities or for rapid prototyping of geometrically complex sandwich

components.

In order to prove the viability of the process, Maley implemented mouldless VARTM on the full

scale fuselage section of the GeoSurv II UAS [3]. The outcome showed a need for improvement

in the process robustness, repeatability, and tolerances. The following sections provide a brief


overview of the GeoSurv II UAS followed by the summary of thesis objectives, thesis

organization and contributions.

1.3. Overview of the GeoSurv II UAS

GeoSurv II is an all-composite UAS, currently being developed at Carleton University as part of a

fourth year undergraduate team project in the Department of Mechanical and Aerospace

Engineering. It is designed to perform multi-purpose geophysical survey missions including high

resolution magnetic surveys that are of particular interest to the industrial partner SGL, an

Ottawa based company specializing in airborne geomagnetic, gravimetric and radiometric

surveys around the world.

SGL currently flies airborne surveys using conventional manned fixed-wing aircraft and

helicopters. A survey mission requires a minimum crew of four people: two pilots, a

geophysicist and an aircraft maintenance engineer. The GeoSurv II is designed to be controlled

autonomously with an autopilot featuring an on-board obstacle detection and avoidance

system. This will enable geophysical surveys to be executed with two operators: a geophysicist

and an aircraft maintenance engineer. The autonomous nature of the GeoSurv II will, in future,

allow the execution of multiple aircraft operations using a single operator, thereby increasing

the quantity of the data collected at reduced operational costs.

The composite airframe design of the GeoSurv II reduces undesirable magnetic noise within

near proximity of the magnetometers mounted on the aircraft wing tips. Benefits of composite

airframe construction coupled with the GeoSurv ll's ability to fly slower and closer to the

ground will substantially improve the quality of the acquired data at much lower capital and


operational costs compared to its manned counterparts. Figure 1.1 shows the current

configuration and specifications of the GeoSurv II [3].

v*& •Wing span 16 ft. 11 Length 14 ft. 11 Height 3 ft. •200lbMTOW •60/100 kts cruise speed •30 hp propeller engine with pusher configuration •Composite airframe

Figure 1.1: Current configuration and specifications of GeoSurv II UAS (2009/2010) [4]

The 2010/2011 academic year is the 7th year of GeoSurv II development at Carleton University.

Thus far, most of the design and analysis of GeoSurv II has been completed and a working

prototype has been constructed. The prototype (Figure 1.2) is currently being prepared for

initial flight testing.

Figure 1.2: GeoSurv II Prototype Assembly (February 2010)

Low-cost composite airframe research is one of five critical areas, which required more

advanced graduate level research during the course of the GeoSurv II development. Other

graduate research topics include autonomous operation, obstacle detection and avoidance,

flight control actuating systems with low magnetic signature and geomagnetic data acquisition.


1.4. Thesis Objectives and Organization

This thesis work focuses on the design and manufacturing of geometrically complex, foam-core

composite sandwich structures, by means of "mouldless" LCM method. The thesis objective is

to develop a robust, low-cost mouldless LCM method for foam-core composite sandwich

structures and utilize the method to manufacture a full scale GeoSurv II fuselage.

In this work, the design of GeoSurv II fuselage and previously implemented mouldless VARTM

methods were studied to identify critical areas that require improvement. Based on this study,

design and manufacturing objectives were proposed for the next generation fuselage. Then,

state-of-the-art LCM processes found in literature were reviewed, which identified Closed

Cavity Bag Moulding (CCBM) as a potential alternative to VARTM for mouldless manufacturing.

A series of experiments was carried out using CCBM process to evaluate its feasibility for

mouldless manufacturing, including investigations of various bag sealing and resin distribution

methods. Feasibility of CCBM for mouldless manufacturing was assessed through a Process

Value Analysis (PVA). Appropriate materials were selected for fuselage manufacturing through

the mouldless CCBM method. Subsequently, the fuselage was redesigned for improved

manufacturability. The new design was optimized using a Finite Element Analysis (FEA)

implemented in Abaqus. The FEA predictions were verified by testing demonstrator coupons at

specific loading conditions. Finally, a full scale GeoSurv II fuselage was manufactured to

demonstrate the viability of the developed process.


This work is organized into the following chapters:

> Chapter 2: Overview of the GeoSurv II Fuselage The design of the GeoSurv II fuselage is discussed. A mouldless VARTM method previously implemented on the fuselage prototype is presented. A post-process assessment is also provided in conjunction with the design and manufacturing objectives for the new fuselage.

> Chapter 3: Review of LCM Processes A summary of the available LCM technologies is presented. Various low cost LCM processes are assessed to select the most suitable process for mouldless manufacturing.

> Chapter 4: Mouldless CCBM Process Development A series of experiments carried out using the chosen CCBM method from Chapter 1 are discussed. Various bag sealing methods and infusion strategies are evaluated to determine the feasibility of CCBM for mouldless manufacturing. Finally, PVA carried out to determine the most feasible mouldless manufacturing method is presented.

> Chapter 5: Fuselage Material Selection State-of-the-art structural sandwich core materials are discussed and suitable core material for mouldless manufacturing is selected. Material choices for the matrix, reinforcement and inserts are also presented along with the rationale for selection.

> Chapter 6: Fuselage Redesign The design changes made to the GeoSurv II fuselage to improve its manufacturability are described. An improved GeoSurv II fuselage model is presented.

> Chapter 7: Design Optimization An FEA carried out on the new fuselage design, to optimize the composite layup is discussed. A weight estimate for the optimized structure is provided. Finally, an experimental verification of the FEA results using demonstrator test coupons is presented.

> Chapter 8: Fuselgge Manufacturing Manufacturing of a full scale GeoSurv II fuselage by an improved mouldless CCBM method is discussed.

> Chapter 9: Manufacturing Results The outcomes of the fuselage manufacturing by mouldless CCBM method are discussed. The results are assessed against the design and manufacturing goals.

> Chapter 10: Conclusions Conclusions drawn from this research work are summarized. Recommendations and future work are discussed.


1.5. Contributions

In this research, state-of-the-art LCM processes currently used for manufacturing composite

components were studied and the processes applicable for mouldless manufacturing were

identified. An in-depth assessment of the applicable processes showed that CCBM should be

considered further for mouldless manufacturing.

CCBM is a relatively new process, currently popular in the marine industry. This resin infusion

process uses silicone based elastomeric vacuum bags that are form-fitted to the shape of the

part. CCBM infusion offers good vacuum integrity and better surface finish compared to

traditional VARTM infusion using disposable vacuum bags. However, current CCBM methods

are intended to be used for manufacturing with rigid moulds and hence can be expensive at the

outset [4]. As part of this research, a series of experiments were carried out to develop CCBM

techniques suitable for mouldless manufacturing. A PVA was used to assess the process

variants and select the most suitable technique (section 4.3) for mouldless manufacturing. This

research laid the foundation for further development of mouldless CCBM methods.

The use of Design for Manufacturing (DFM) principles and FEA for continuous design

improvement was demonstrated. An effective methodology for creating bolted joints in foam

core composite sandwich structures was developed and demonstrated.

DFM principles were applied to improve the manufacturability of the GeoSurv II fuselage. The

new design was optimized using a simplified FEA carried out using Abaqus. An experimental

study was conducted to verify selected FEA results. Approximately 36% weight savings were


achieved with this design optimization work. The Abaqus FE model of the fuselage can later be

expanded to include more realistic material property formulations and dynamic loading.

This thesis has contributed to the development of a unique mouldless manufacturing technique

for foam-core composite components, using an improved CCBM method featuring Channel-ln-

Bag (CIB) infusion. A full scale GeoSurv II fuselage has been manufactured in a single infusion

step, using this mouldless manufacturing method. This is a low cost LCM process suitable for

producing complex geometry foam-core composite components with good part quality and

process repeatability. The process is economical for part quantities below 10 and hence is

beneficial for low to medium volume production runs and prototype development. This

manufacturing method should be considered for small aircraft and UASs featuring foam-core

composite designs.

This research has successfully addressed the major aspects of low cost composite structures

including structural design with material selection, structural optimization, DFM, PVA,

manufacturing process planning and development.


CHAPTER 2. OVERVIEW OF THE GEOSURV II FUSELAGE

This chapter provides an overview of the GeoSurv II UAS fuselage and describes the mouldless

VARTM method previously implemented for manufacturing the first fuselage prototype. A post

process assessment is also provided followed by the design and manufacturing objectives for

the new fuselage.

2.1. Design of the GeoSurv II fuselage

Major components of the current GeoSurv II fuselage assembly are shown in Figure 2.1. It

consists of a structural main frame ("H-frame"), upper and lower access hatches, and a

nosecone. The access hatches and the nosecone contribute mainly to aerodynamic drag

reduction and are subjected to relatively low structural loads, with the exception of the lower

access hatch, which carries 20 lbs of geomagnetic survey payload.

Figure 2.1: Main sections of the GeoSurv II fuselage [3]

CHAPTER 2. OVERVIEW OF THE GEOSURV II FUSELAGE 10

The structural main frame features an "H" cross-section with foam core carbon-epoxy sandwich

construction. Figure 2.2 shows the cross-section of the main frame and highlights critical areas

through which the in-flight and the landing loads are transferred into the fuselage. The in-flight

bending stresses from the wings are transferred through the carry-through spar, while the

aerodynamic moments are transferred through two shear pins located forward and aft of the

spar.

The front and rear bulkheads carry the nose landing gear loads and the engine loads

respectively. Two bolts directly above the carry-through spar location transfer loads from the

main landing gear into the fuselage walls. An airfoil shaped fairing is incorporated in the design

to provide smooth transition from the fuselage to the wings and thus minimize interference

drag.

Rear bulkhead: Engine loads

Main iandinggearattachment Aft shear pin location

Carry through sparlocation

Forward shear pin location

Front bulkhead: Nose Landing gear/Air data boom loads

Figure 2.2: Cross-section of the fuselage structural H frame [3]


2.2. Manufacturing of the Fuselage Prototype

A prototype of the fuselage main frame was manufactured using a mouldless VARTM method

developed by Maley [3]. Mouldless VARTM is a low cost processing method applicable for

composite sandwich components, in which the foam core is used as the mould during

manufacturing. This manufacturing method includes three primary steps: core preparation,

fabric layup and infusion, all of which are shown Figure 2.3. In step 1 "Core Preparation", the

required components were machined from Extruded Poly Styrene (EPS) insulation foam1. These

foam components were assembled using Airtac 2 spray adhesive. The foam assembly was then

coated with a thin layer of West System 105 epoxy resin and finished with West System 407 (a

low density fairing filler), both of which were used to increase the core stiffness. In step 2:

"Fabric Layup", the prepared foam component (the core assembly) was laid-up with Hexcel -

AGP 370 5 HS fabric at [(0/90)°, (+/- 45)°, Foam Core]SYM orientation.

Figure 2.3: Sequence of the main steps employed in mouldless VARTM

EPS (pink) foam was used in the GeoSurv II prototype, primarily to reduce cost. Using an aerospace grade structural core material would aid in better dimensional tolerances and may result in significant weight saving, as it eliminates the need to seal the foam prior to infusion.


In step 3: "Infusion", optimum locations for resin inlets and outlets were determined through

the use of Liquid Injection Moulding Simulation (LIMS) software [3]. In this process, 1-D

permeability values were first measured for the fuselage layup. These permeability values were

used with the LIMS software to simulate the resin flow during infusion. Simulation results were

used to determine the optimal locations of both resin inlets and outlets, to improve infusion

quality and to reduce infusion time. The infusion setup was prepared with the resin inlet and

outlet lines placed at the locations suggested by the flow simulation. The fuselage was then

manufactured in a single infusion step. A detailed description of this process simulation and

manufacturing method can be found in [3].

2.3. Post Process Assessment of the GeoSurv II Fuselage

The prototype fuselage manufactured by mouldless VARTM, is shown in Figure 2.4. One of the

major problems encountered was the occurrence of vacuum bag leaks during resin infusion.

The leaks resulted in small air pockets at the corners of the cured fuselage. Furthermore, the

foam core was distorted under the applied vacuum pressure, leading to an average deviation

from flatness of 0.15 in. over the side wall (length: 44 in.) and up to 0.31 in. deviation from the

target dimension at several locations [3]. The manufactured component was sanded and filled

to reduce the dimensional variations and improve the surface finish. The access hatches and

several other components that interface with the fuselage main frame were modified to mate

with the distorted shape of the as-manufactured fuselage. Additionally, extensive labour was

required to complete the fuselage assembly, shown in Figure 2.5. This included drilling holes

and bonding rigid inserts into the holes to facilitate load transfer into the structure at various

joints, manufacturing and bonding of the nose cone bridges, bay separator panel and mounting


brackets for the mission avionics rack. Thus, the cost of labour required to carry the fuselage

into a finished stage outweighed the economic advantages of mouldless VARTM. This indicated

a need for improvement to both the structural design and the manufacturing process.

Figure 2.4: Fuselage main frame manufactured by mouldless VARTM

Upper nosecone bridge

Holes/inserts for mountingNLG and air data boom

Wooden brackets to mountthe nosecone

Figure 2.5: Fuselage prepared for assembly


2.4. Design and Manufacturing Objectives for the New Fuselage

Enhanced with the appropriate selection of new materials and an improved processing method,

the new fuselage shall replace the prototype in its present configuration. The new fuselage shall

be manufactured using a low-cost, mouldless LCM method. The manufactured component shall

comply with the GeoSurv II airframe requirements, listed in Table 2.1. The design and

manufacturing objectives of the new fuselage are given in Table 2.2. The work-flow involved in

the development of the new fuselage is shown in Figure 2.6.

Table 2.1: Airframe requirements for the GeoSurv II UAS [5]

GeoSurv II UAS Airframe Requirements > Materials used in the primary and secondary structures shall be non-ferrous to the greatest extent

possible, to minimize magnetic noise. > Light weight modular airframe. > The structure shall be robust and relatively inexpensive to manufacture and maintain. > The UAS should operate within a temperature range of -49 to 104 °F (-45 to 40 °C) > The UAS airframe should comply with Canadian Aviation Regulation, Part V, 523-VLA (Very Light

Aircraft): Fire/smoke resistance, moisture, chemical corrosion resistance.

Table 2.2: Design and manufacturing objectives for the new fuselage

Re-design Objectives > Improve the design to reduce weight and

increase structural integrity. > Apply effective Design for Manufacturing

(DFM) and Integrated Manufacturing (IM) principles to facilitate near-net-shape manufacturing by mouldless LCM method.

Manufacturing Objectives > Improve part quality and tolerances (near-net-

shape manufacturing). > Improve process repeatability.

Note: The first prototype of the GeoSurv II fuselage developed by Maley [3] is referred to as the

current fuselage, and the fuselage developed as part of this research is referred to as the new

fuselage throughout the rest of this thesis.


Problem Definition Review of the current fuselage

design/manufacturing (Chapter2)

Literature Review and Process Selection

(Chapter 3)

Process Optimization and Feasibility Studies

(Chapter4)

Material Selection (Chapters)

Design Optimization (Chapter 7)

New Fuselage Design (Chapter6)

New Fuselage Manufacturing (Chapters: 8,9)

Figure 2.6: Work-flow diagram of the new fuselage development


CHAPTER 3. REVIEW OF LCM PROCESSES

This chapter provides an introduction to LCM processes and discusses several LCM variants

used in the industry, for manufacturing composite components. Following the review, low cost

LCM processes applicable for mouldless manufacturing were assessed to identify the most

suitable process.

3.1. LCM Processes

LCM refers to a family of processes, in which dry fibrous reinforcement materials placed in a

mould cavity are impregnated with a liquid polymer matrix (typically thermoset resins) under a

forcing pressure gradient (vacuum pressure, positive pressure or both). The resin is then

allowed to cure; the part is demoulded and subjected to finishing operations as required, to

create the final product. LCM processes can be implemented using many low-cost tooling and

part constituents for a wide range of part size, complexity and production quantity.

Characterized by a closed mould setup, LCM processes further offer an operator-friendly

manufacturing environment, good fibre orientation control and improved process repeatability.

Along with these benefits there exist some inevitable drawbacks, such as relatively complex

processing steps and difficult quality control with in-house resin mixing [6-9]. The potential

advantages and disadvantages of LCM processes compared to other traditional composite

manufacturing methods are summarized in Table 3.1.


Table 3.1: Potential advantages of disadvantages of LCM processes [6-10]

Advantages > Low material, initial equipment and recurring

costs compared to prepreg-autoclave method > Many material choices for reinforcement,

matrix and core > No size restrictions: Integrated manufacturing

of large structural components is possible > Low cost LCM methods yield more consistent

part quality than wet layup and vacuum bagging, at similar costs

> Greater flexibility than any other processing method

> Thick laminates can be processed easier than with wet layup

> Produced laminates have uniform microstructure and minimum void content compared to wet layup

> Sandwich constructions can be produced in single infusion process

> Characterized by closed-mould operations, the LCM methods reduce volatile organic compounds (VOC) emission from resin by more than 90% as compared to open-mould hand lamination methods

Disadvantages > Requirement for relatively complex/different

skills than wet layup > Requirement for resins with low viscosity may

compromise the thermal and mechanical properties of some polyester/vinyl based laminates

> Low fibre volume fraction compared to prepreg-autoclave method (Except for RTM type processes)

> Air leaks in the tool or bag and uneven resin flow may result in resin rich/starved regions and in turn cause expensive scrap parts

> With the exception of RTM type processes and LRTM, most other LCM processes produce parts with moulded finish on one side.

> Difficult to implement on the honeycomb cores

Attracted by the potential cost savings and process flexibility, manufacturers are increasingly

researching and developing LCM techniques for producing their components. This has led to

the evolution of many LCM process variants over the last few decades. Despite the complexities

associated with specific process variants, all LCM processes share several distinctive features

[7,8]:

> A resin delivery system > A mould setup equipped with appropriate clamping and manipulation devices > A reinforcement handling system (i.e. fibre preforms, sheet/bulk mounding compounds) > A strategy for air displacement or evacuation and resin supply


The functionality of these systems is determined by the scale of the part and complexity of the

manufacturing operation.

Based on the forcing pressure gradient used to introduce the resin into the mould cavity, the

LCM process variants can be grouped into two categories: Resin Transfer Moulding (RTM) and

Vacuum Assisted Resin Transfer Moulding (VARTM). The former uses positive pressure to inject

the resin into the mould cavity, while the latter uses a negative pressure gradient to draw the

resin into the mould cavity. Figure 3.1 shows a family of state of the art LCM processes

developed from RTM and VARTM. Unique features of each of these processes are described in

the following sections.


Liquid Composite

Moulding (LCM) Processes

Resin Transfer Moulding

(RTM) & Structural

Reaction Injection

Moulding (SRIM)

Vacuum Assisted Resin

Transfer Moulding

(VARTM)/VAP/VIMP

- »

Seemann Composites Resin Transfer Moulding (SCRIMP™)

Vacuum Assisted Process (VAP®)

Fast Remotely Actuated Channelling (FASTRAC)

Controlled Atmospheric Pressure Resin Infusion (CAPRI)

Double Bag VARTM

Advanced VARTM (A-VARTM)

Single Line Injection (SLI)

High Performance VARTM/RTM (Hyper-VARTM™, Hyper-RTM™)

Vibration Assisted Liquid Composite Moulding

Closed Cavity Bag Moulding (CCBM)

Co-Injection Resin Transfer Moulding (CIRTM)

Euro Composites® Honeycomb Liquid Moulding (EC-HLM)

VacFlo® Process

Light RTM

Resin Infusion between Double Flexible Tooling (RIDFT)

Flexible Injection

Resin Film Infusion (RFI)

Semi-Preg Infusion

Figure 3.1: A family of state of the art LCM processes

3.1.1. Resin Transfer Moulding (RTM)

RTM is best known for its ability to fabricate large, complex parts to near-net shape with

excellent surface finish on both sides. A schematic of a basic RTM setup is shown in Figure 3.2.

In this process, dry fibre preform is laid-up into a two part rigid mould, and resin is injected into

the mould cavity until the preform is fully saturated. The part is then cured and removed from

the mould. RTM produces near autoclave quality parts with void content <1%, at very short

cycle times (typically 3-5 minutes). However, it requires matched, leak proof moulds capable of


withstanding the injection pressures up to 1500 psi. Designing and manufacturing RTM moulds

require high initial costs, making RTM economical for large production quantities, typically

above 1000 units [10,11].

Resin Injection Port

ClampingPressure

Figure 3.2: Schematic of the Basic RTM setup

3.1.2. Structural Reaction Injection Moulding (SRIM)

SRIM is an LCM process evolved from the family of Injection Moulding (IM) processes. IM is a

process whereby liquid thermoplastic or partially polymerized thermoset resins are injected

into mould cavity to produce non-reinforced polymeric components. This process is typically

carried out at elevated temperatures and pressures for rapid part production in large

quantities. A further development to IM is the Reaction Injection Moulding (RIM): a process in

which two part thermoset resins are mixed and injected into the mould cavity to produce the

final part. The processes take their name from the chemical reaction that takes place within the

machine [8]. When short fibre reinforcements are included into the RIM, the process is called

Reinforced Reaction Injection Moulding (RRIM). SRIM is a further development to RRIM,

whereby preformed fibre mat reinforcements are impregnated with liquid resin to produce the


final part. The sequence of steps employed in SRIM is shown in Figure 3.3. This process is

characterized by mould fill times less than 1 minute and pressures greater than 1500 psi.

Figure 3.3: SSRIM process [8]

The RTM and SRIM equipments, though relatively inexpensive compared to an autoclave

infrastructure, require somewhat high initial costs. Hence, they are efficient for high volume

production runs. Since the primary focus of this research is low volume production by LCM

process, RTM and its process variants were not studied in detail. Substantial reviews of RTM

type processes can be found in references [8 -11].

3.1.3. Vacuum Assisted Resin Transfer Moulding (VARTM)

The need for an LCM process for economic part production in low production quantities has

resulted in the development of Vacuum Assisted Resin Transfer Moulding (VARTM). VARTM is a

variation of RTM, in which resin is drawn into the preform under vacuum pressure, rather than

being injected at positive pressure. A schematic of the VARTM process is shown in Figure 3.4.


This process begins with preparation of a rigid mould and layup of dry reinforcement in the

mould cavity. Then, a flexible vacuum bag is placed over the reinforcement and is sealed

against the mould using sealant tape. A vacuum tight chamber is created inside the bag by

connecting the outlet to a vacuum pump. Resin is then introduced into the mould via an inlet,

which is typically on the opposite side of the outlet. The reinforcement is impregnated by the

resin due to the pressure differential between the inlet and the outlet ports. When the

reinforcement is fully impregnated, the resin is left to cure under the vacuum pressure to

create the final part. In order to reduce the flow resistance and hence the filling time, a highly

permeable distribution medium is introduced between the top layer of the reinforcement and

the vacuum bag. During infusion, the resin distribution medium lifts the bag surface away from

the preform, leaving a highly permeable path for resin to travel. Resin is distributed

preferentially through the surface of the flow media and simultaneously through thickness of

the preform [3,12,13].

Figure 3.4: Schematic of basic VARTM process

VARTM is a low-cost process capable of manufacturing large-scale complex geometry

composite parts with good surface finish and excellent dimensional tolerances. This processing


method eliminates the need for expensive moulds required by the RTM and allows replacement

of the upper mould with a flexible polymeric bag. Hence, VARTM is sometimes referred to as

Resin Infusion under Flexible Tooling or RIFT. Other generic terms given for VARTM include

Vacuum Infusion (VI), Vacuum Infusion Moulding (VIM) and Vacuum Assisted Resin Infusion

Moulding (VARI/VARIM) [6].

Since the development of VARTM processes, researchers and manufacturers have focused on

improving the resin infusion characteristics and processing methods to further reduce costs and

improve part quality. This has led to the development of many materials and improved VARTM

techniques for tooling setup, resin distribution and complete resin impregnation. Some

important developments are outlined in Table 3.2. Many others combined new materials with

innovative process modifications, while attempting to bring out solutions for existing challenges

with conventional VARTM methods. The following sections consider some important VARTM

process variants that originated from research and development work carried out worldwide,

over the last two decades.


Table 3.2: New materials and technologies developed to improve VARTM process

Research Area Matrix/Core/ Reinforcement Materials

Resin Distribution

Flow Simulation

Tooling/ Vacuum Bagging Supplies

New Developments > Matrix: Epoxy, polyester and vinyl-ester resins available with viscosities below 200

cps and wide range of working times and curing temperatures. Suppliers Include: PTM&W Industries Inc.-Epoxy-N-Fusion®, Gougeon Brothers Inc-West System®, Tri-Tex Co. Inc. - East System®, Polymer Tooling Systems Inc. - Renlnfusion® Epoxies, Endurance Technologies Inc. -Endurance Epoxy Systems, Applied Poleramics Inc.

> Woven Fabric Reinforcement: Carbon, Kevlar, Glass, Aramid or hybrid fabrics available in different densities and variety of fabric styles for improved drapability.

> Reinforcements are also available in roving, mat, knit and 3-D braided forms [6,10]. > Non-Crimp (NC) fabrics with or without tackifiers to facilitate easy layup [6]. > Suppliers of reinforcement materials include The Saertex Group, Metyx Composites,

Hexcel Corporation, Polynova™ Composites., Vectorply Corporation, and JB Martin > Variety of core materials available for infusion (Chapter 5). > Low cost polyethylene based distribution media available for high and low

temperature processing. (Airtech: Resinflow 60, Resinflow 90 HT, Delstar Naltex®) > High Permeable Layers (HPL) for faster and quality infusion. Examples include:

Airtech N4, Airtech N10, Richmond A3000, Colbond EnkaFusion. > Reusable breather and resin distribution medium: Airtech Breatherflow 60. > Inter-laminar infusion for processing thick laminates by VARTM: Using 3-D spacer

fabrics engineered as distribution media or using Lantor Soric® flexible core. > Channel-ln-Bag (CIB) Infusion: Resin conduits integrated into reusable silicone

vacuum bags for resin distribution [14]. > Channel-ln-Core (CIC) Infusion: Resin channels grooved into foam cores to

distribute the resin into the part [15]. > Liquid Injection Moulding Simulation (LIMS): VARTM/RTM flow simulation software,

developed by the University of Delaware- Centre for Composite Materials. > PAM-RTM 2004 LCM simulation software: Flow simulation and process optimization

software for RTM and VARTM type processes developed by the ESI Group. > Simulation based Liquid Injection Control (SLIC): A software package that integrates

flow simulation, setup optimization and process automation and control to provide optimum LCM design and manufacturing solutions; developed at the University of Delaware- Centre for Composite Materials.

> RTM-Worx: Flow simulation software capable of simulating and optimizing RTM/VARTM type processes, developed and licensed by Polyworx Corporation.

> Disposable/reusable vacuum bags available in various levels of conformability (i.e. Econolon®, Stretcholon®, Multibag®).

> Peelplies: Polyester or nylon based, available for low and high temperature processing; silicone or Teflon coated peelplies are also available for easier release.

> Resin infusion lines: Disposable spiral tubing and reusable Omega Flow Lines. > Spray adhesives for holding fabrics together during layup (i.e. Airtac 2, Econotac). > Airtech Fusiontac: A preform tackifier compatible with polyester/vinyl ester resins > Airtech TackStrip: An adhesive coated mesh for holding fabrics during layup- This is

an alternative to spray adhesives developed to prevent laminate contamination. > Suppliers of VARTM tooling materials include Airtech International Inc., Aerovac

Systems Inc., Advanced Composite Materials Inc., and Torr Technologies Inc.


3.1.4. Seemann Composite Resin Infusion Moulding Process (SCRIMP™)

SCRIMP™ is a VARTM process variant developed in late 1980's by Bill Seemann and patented in

1990 by TPI technologies. SCRIMP™ was developed to reduce infusion times for large and

complex parts and thus improve the overall production rate. The sequence of steps involved in

SCRIMP™ is similar to VARTM except that TPI holds the patent for the particular flow media

and flow process, thus being able to charge for using this particular infusion method [3,16].

3.1.5. Vacuum Assisted Process (VAP®)

The Vacuum Assisted Process (VAP*) was developed by EADS Deutschland and is protected by

several worldwide patents for low cost manufacturing of primary aircraft structures. This

process is characterized by inclusion of an additional Gore Composite Manufacturing (GCM)

membrane, to the conventional VARTM setup, as shown in Figure 3.5. The GCM membrane is a

selectively permeable material that acts as a barrier to resin but is highly permeable to air.

Breather / Flow media Chamber 1 fVaeuuml V a c u u m f o j | Row media Membrane

Vacuum „JU

L! &&

Resin Injection fine

CFC- Preform

Vacuum r

Tooling Sealant Tape

lama si".

*S?SS?^S;£H-J

A

Figure 3.5: Schematic of the EADS VAP8 before (top) and after (bottom) infusion [10]


Adding this membrane over the distribution medium of a conventional VARTM setup leads to

application of more uniform vacuum pressure on the laminate throughout the infusion process.

This is claimed to result in uniform part thickness, low void content, minimum potential for dry

spot formation and improved overall part quality [10,17].

3.1.6. Fast Remotely Actuated Channelling (FASTRAC)

The FASTRAC process was first developed at the United States Army Research Laboratory (US-

ARL). Instead of using a flow medium to accelerate the resin flow, this innovative process uses a

low cost rigid or flexible tooling in a double vacuum bag setup, in order to create preferential

resin flow channels along the part. A basic FASTRAC infusion setup is shown in Figure 3.6.

Secondary FASTRAC vacuum bag, FASTRAC Non-contacting Tool

Figure 3.6: Schematic of the FASTRAC process [18]

In this process, the preform is placed under a primary vacuum bag and kept under full vacuum

pressure (close to -30 inHg). Then a secondary vacuum bag featuring a non-contacting FASTRAC

tool is placed over this primary vacuum bag and sealed to the open mould surface. Vacuum is


drawn inside this secondary vacuum bag to keep the FASTRAC tool under full vacuum pressure.

To create the highly permeable resin channels, vacuum pressure inside the primary vacuum bag

is released to relax the bag for a moment and then drawn again to full vacuum pressure. This

allows the primary vacuum bag to form into the FASTRAC tool and create highly permeable

channels that preferentially distribute the resin over the part. Once the part is fully saturated

with resin, the FASTRAC layer can be removed and the part can be cured [18].

3.1.7. Controlled Atmospheric Pressure Resin Infusion (CAPRI)

CAPRI is a VARTM process variant developed and patented by the Boeing Corporation. As

shown schematically in Figure 3.7, CAPRI process includes two adjustments to conventional

VARTM methods, aimed at improving the fibre volume fraction and reducing the thickness

gradient of the final part.

Vacuum Distribution b a 9 media Preform Resin inlet

Figure 3.7: Schematic of the CAPRI process [19]

First, the dry preform is subjected to cyclic debulking under vacuum pressure, after bagging but

prior to infusion. This is said to improve the fibre nesting, reduce the thickness of the laminate


and hence, improve the overall fibre volume fraction. Additionally, the infusion bucket

containing the resin is kept under partial vacuum using a secondary vacuum pump, while the

outlet lines are kept under full vacuum. The partial vacuum applied to the resin chamber,

though increasing the process complexity and infusion times, reduces the thickness gradient

observed in conventional VARTM processing.

3.1.8. Double Bag VARTM

The Double bag VARTM process is an improved version of conventional VARTM developed and

patented by the Boeing Company, to solve the bag relaxation and thickness gradient formation

associated with conventional VARTM. This process adds a second vacuum bag separated by a

layer of breather cloth to the conventional VARTM setup, as shown in Figure 3.8. The second

vacuum bag applied around the part acts as a caul plate preventing the inner bag from relaxing

behind the flow front. This added bag also restrains the first bag from stretching during and

after the infusion, which helps in maintaining the vacuum integrity. The double bagging process

is expected to reduce the thickness gradient, improve fibre volume fraction and hence the

overall part quality [12].

To vacuum port

Resin feed/

'"'et line Breather cloth

A\..: /_(*>_ i—r —ra^" 1^^™"^^^T. ^^^' —r

Outlet Line: to vacuum port

Figure 3.8: Schematic of double bag VARTM

Outer vacuum bag


3.1.9. Advanced VARTM (A-VARTM)

The A-VARTM process is a variation of VARTM developed jointly by Mitsubishi Heavy Industries,

Ltd. and Toray Industries, Inc. for fabrication of aircraft primary structures. This process

combines traditional VARTM methods with the addition of preform hot-compaction prior to

infusion and resin bleed-off after infusion, in order to obtain nearly 60% fibre volume fraction.

Enhanced unidirectional properties are attained with the use of advanced Non-Crimp Woven

(NCW) fabrics (Figure 3.9). Primary aircraft structures are manufactured as integral components

by applying co-bonding and advanced pre-forming techniques. Figure 3.10 shows an outline of

the A-VARTM process [20].

Figure 3.9: Outline of NCW fabric [20]


Carbon fiber | Preparation of tools

( Weaving )

NCW

G I

Hot compaction

( Resin tnfusion.'bleed

( First cunng )

( Post curing J

VaRTM composite

Base resin Hardener X

/ Measured N /Measured I volume J I volume

f De-foam )

I f Mixing j

Figure 3.10: Sequence of steps in A-VARTM process [20]

3.1.10. Single Line Injection (SLI)

SLI is a cost-optimized Liquid Resin Injection (LRI) process to manufacture high quality

composite components in the autoclave. Developed at the German Aerospace Centre (DLR)

Institute for Structural Mechanics, the SLI process combines the advantages of the cost efficient

materials and liquid resin injection processes to manufacture high quality parts in an autoclave.

A schematic illustration of SLI is shown in Figure 3.11 and Figure 3.12.

pressure reducing valve

resin container

vacuum system

resin transfer line

fibre preform

single-sided tool

Figure 3.11: Illustration of SLI process [21]


P ~ P Pressure P S P 1 autoclave • injection r l g M u l g ' autoclave/ • injection

ffl__ ____— Reducing-Valve

Injection phase

K>

p p ' injection injection

autoclave . ^ k / \ _ _ / ^ _ _

Adjustment of fibre content

Figure 3.12: Pressure distribution during and after resin injection in SLI process [21]

SLI simplifies the traditional VARTM and RTM setups by utilizing a single resin transfer line for

both evacuation of preform and resin injection. This is accomplished by adjusting the pressure

accordingly at different stages as shown in Figure 3.12. Prior to resin injection, the autoclave is

set at pressure level higher than full vacuum pressure while the mould cavity with fibrous

reinforcement is kept under vacuum pressure. Then, resin is injected into the part at a pressure

level equal to that of the autoclave pressure in order to impregnate the part. Following

complete resin impregnation, the pressure inside the autoclave is increased until the desired

fibre volume fraction (typically 60%) is attained and excess resin is drawn out of the resin feed

line with the assistance of the vacuum system. Depending on the part geometry and size, the

placement of the resin feed line can be optimized [21].

3.1.11. High Performance VARTM (Hyper-VARTM™/ Hyper-RTM™)

Hyper-RTM™ and Hyper-VARTM™ are innovative LCM technologies patented by V Systems

Composites Inc. (VSC), San Diego, CA (US). These technologies incorporate the resin

distribution system into the process' proprietary tooling technology, which allows controlled

resin propagation along both in-plane and out-of-plane directions relative to the tool surface.


Their tooling incorporates the so-called 'port runner devices' that are claimed to be universal,

modular, reusable and provide a non-directional, high permeable system to facilitate the resin

infusion. The result as claimed in the patent (US 6,964,561 B2, Nov 15, 2005), is consistent and

high quality infusion with minimized setup labour and reduced potential for rework or scrap

[22].

3.1.12. Vibration Assisted Liquid Composite Moulding

In this LCM method, either the mould or the resin stream is oscillated using electromagnetic

shakers or motors during the infusion. Such vibrations assisted methods have been shown to

reduce void contents in the final part, enhance flow rates and improve the fibre wetting [23].

3.1.13. Closed Cavity Bag Moulding (CCBM)

CCBM is a relatively new process for low cost manufacturing of FRP composites, patented by

Arctek Inc. It utilizes a silicone based elastomeric material to manufacture flexible vacuum bags

that are form fitted to the shape of the mould [24]. Once the bag is manufactured, the process

follows conventional VARTM methods, as shown in Figure 3.13.

{Photo courtesy of Progress Plastics & Compounds Company Mlsslssougo. ON. Canada, 2008}

Figure 3.13: Typical CCBM procedure


CCBM bags offer good vacuum integrity and better surface finish compared to traditional

disposable vacuum bags. A properly made CCBM bag can last up to 1000 manufacturing cycles,

leading to potentially significant cost savings in material and labour for small to medium

production runs [24].

CCBM systems are typically supplied in sprayable or brushable forms. Sprayable CCBM systems

generally involve an initial equipment cost, hence are somewhat expensive at the outset

compared to the brushable CCBM systems that only require brushes and squeegees for

manufacturing the bag. Commercially available CCBM type systems include, SWORL™,

Sprayomer Elastomer™, Airtech Multibag™ and Vacuspray™.

3.1.14. Co-Injection Resin Transfer Moulding (CIRTM)

Co-Injection Resin Transfer Moulding (CIRTM) is an LCM method developed jointly by University

of Delaware's Center for Composite Materials (UD-CCM) and Army Research Laboratory (ARL).

In this process, manufacturing of hybrid composites in a single step is achieved by simultaneous

injection of multiple resins into a multi-layer preform. This method can be utilized with

conventional VARTM setup as shown in Figure 3.14.


phenolic J2027

vinyl ester 411-350

flow direction

vacuum line

SOO<>0<i<X>iJ<M^O<Vg»g o /

distribution media

separation layer

window

mold surface

vacuum bag

release ply

distribution media

Figure 3.14: Schematic of the CIRTM method [25]

The hybrid structure is facilitated using a separation medium, which isolates the individual

resins during infusion but forms a good structural bond upon cure. CIRTM results in substantial

cost savings by the eliminating the need for multi-step processing and secondary bonding when

manufacturing a hybrid composite structure [25].

3.1.15. Euro Composites® Honeycomb Liquid Moulding (EC-HLM)

EC-HLM is a process in which sandwich constructions with honeycomb cores can be produced

using the LCM method, without filling the honeycomb cells with resin. This is achieved by

integrating a unique barrier layer into the fabric layup process as shown in Figure 3.15. The part

is then vacuum bagged as shown in Figure 3.16. In an integrated infusion process, pre-curing of


the barrier-core-bond takes place in an oven first. This is followed by resin infusion and final

cure. The process is claimed to produce good quality, out-of-autoclave honeycomb structures

[26].

Barrier

Dry fabric/ fiber pack

material: Honeycomb

Figure 3.15: Details of the part during layup [26]

T/"

Resin Vacuum pump

Resin inlet vent

Vacuum Bag

Draining medium

Part in mold after layup Mold

Figure 3.16: Schematic of the EC-HLM process [26]


3.1.16. VacFlo® Process

The VacFlo® process was developed and patented by Scott Bader Company Ltd., for

manufacturing vacuum infused parts with mould-finished surfaces on both sides. This is

accomplished with the use of matched two-part moulds having a wider flange area that seats a

double-vacuum seal arrangement. During infusion, a first vacuum source is used to draw

vacuum through the flange seal and thus clamp the moulds, while a second vacuum source

facilitates resin flow into the part. The moulds are light weight and typically manufactured using

low cost, fibreglass reinforced laminates. The resulting parts are of same quality as those

produced by VARTM, but show excellent surface finish on both sides [6].

3.1.17. Light RTM

LRTM processes combine certain RTM and VARTM principles to produce vacuum infused parts

with mould-finished surfaces on both sides. This process involves low-cost, two-part moulds

extended with flange-seals, a vacuum source, a low pressure resin injection pump and a

pressure control unit. The vacuum source serves to clamp the moulds and facilitate resin flow

into the part. Resin is introduced into the part using the injection pump, which pushes resin

into the part at pressures that do not exceed the clamping pressure of the mould cavity. In a

typical LRTM setting, the flange seal is clamped with full vacuum pressure (-14.7 psi gauge),

while the mould cavity is maintained at approximately half-vacuum (-7.5 psi gauge). Resin

delivery pressures are constantly controlled during infusion to prevent it from overpowering

the clamping pressure. The push-pull resin infusion method results in complete resin

consolidation of the part in shorter infusion times. This quality infusion comes with a slightly

higher cost than conventional VARTM methods, but the use of low cost moulds and low


pressure equipment makes LRTM relatively inexpensive compared to traditional RTM methods.

Commercial suppliers of controlled LRTM systems include Magnum Venus Plastech (Lite-RTM),

and JHM Technologies (Zero Injection Pressure (ZIP)-RTM) [6,26,27].

3.1.18. Resin Infusion between Double Flexible Tooling (RIDFT)

RIDFT is an innovative LCM process developed to reduce the infusion times associated with

conventional VARTM. Figure 3.17 illustrates the steps involved in the RIDFT. This process

begins with cutting of the dry preform to fit the desired mould shape. The preform is then

placed in-between two flexible silicone membranes and sealed around the edges. Vacuum is

then pulled between the flexible membranes and the preform is infused while remaining in flat

configuration, as shown in Figure 3.17. The impregnated preform is then vacuum-formed to the

shape of the mould by sealing and evacuating the cavity between the bottom silicone layer and

the mould. Finally the part is cured and removed from the mould. Because the preform is flat

during resin infusion, problems associated with wetting out complex geometry components are

eliminated. Since the part stays over the flexible silicone layer at all times, less cleanup and pre-

manufacturing mould preparation is required with each manufacturing cycle. This process is

claimed to produce parts with low void content, at low emissions of Volatile Organic

Compounds (VOCs) and lower tooling costs compared to conventional RTM methods. However,

parts that can be manufactured with this method are of somewhat limited geometric

complexity and size, due to the vacuum forming aspect of this process and the largest available

machine size (10 ft x 15 ft x 4 ft height- Figure 3.18) [29].


,j. ^

Step 1: Load fiber Step 2: Seal Resin Infusion Bag

Step 3: Resin Infusion

L

Step 4: Seal Vacuum Chamber

1 Step S: Vacuum

Form Part Step 6: Demold

Figure 3.17: Schematic of the RIDFT process [10]

Figure 3.18: Industrial RIDFT machine 10 ft x 15 ft x 4 ft [29]


3.1.19. Flexible Injection

Flexible Injection is a new, patent pending LCM technology, developed and currently being

researched at Ecole Polytechnique de Montreal, in collaboration with General Motors (GM) of

Canada Ltd. Flexible Injection can produce parts much faster than other LCM processes without

compromising the part quality. The major steps involved in the Flexible Injection process are

illustrated in Figure 3.19 [30].

The process begins with the placement of fibrous reinforcement into the injection chamber.

Then a flexible membrane is placed over the reinforcement stack and sealed around the mould

(Figure 3.19 - Step 1). Full or partial vacuum may be applied at one extremity of the injection

chamber causing the flexible membrane to compress the reinforcement stack as shown in step

2. The required amount of resin is injected under pressure into the mould cavity, which fills the

fraction of the mould cavity closer to the injection port as shown in step 3. The resin injection is

immediately followed by injection of a non-reactive compaction fluid under pressure, into the

upper mould chamber. This step accelerates the resin flow and impregnation into the fibrous

reinforcement. After complete impregnation of resin into the reinforcement stack, resin vents

are closed and the part is allowed to cure at the desired temperature and pressure. Once the

part is cured, the compaction fluid is drained out of the cavity and the part is demoulded. The

initial process development work demonstrates that the Flexible Injection technology offers

faster and quality resin infusion [30].


Compaction Quid ' outlet

Compaction fluid inttet

Step 1. Mould Setup

A - * *

Step 2. Vacuum Application

< ompartio* Hiiirt Compaction Fluid

Closed

r ^ ^ j Impregnated

I I » . T

Step 3. Resin Injection s t eP 4- Homogeneous compaction of impregnated fibres

Figure 3.19: Flexible Injection Process [30]

3.1.20. Resin Film Infusion (RFI)

RFI uses solid resin film or so called 'prepreg resin' to saturate the dry preform with resin. In the

basic RFI process, the tool (mould) is covered with resin film of required thickness and dry

preform is laid-up on top of this resin film as shown in Figure 3.20. The mould is then bagged

with disposable materials. Subsequently, vacuum or autoclave pressure is applied to the setup,

in conjunction with an appropriate thermal cure cycle. Upon curing, the resin viscosity lowers,

allowing the resin to diffuse into the structural preform. With the RFI setup, the resin only

infuses through thickness, providing considerably shorter infusion times versus traditional LCM


processes. However, RFI requires high temperature rated moulds and at least an oven for

curing, which makes this process somewhat expensive compared to conventional LCM methods

[31].

Bagging film N10 breather

Al tool \

Sealing tape

Bleeder

Perforated film Flashbreaker

— Vacuum

Resin film

Dry preform Cork dam

Airpad caul plate with bleed-out holes

Figure 3.20: Schematic of the RFI process [31]

3.1.21. Semi-Preg Infusion

In this form of resin infusion, reinforcements partially impregnated with resin are used along

with standard vacuum bagging methods and oven curing to produce the final part. These so

called 'sermi-pregs' or out-of-autoclave prepregs are low cost alternative to standard prepregs.

They exhibit shelf life of up to one year at room temperature. Commercially available semi-preg

systems include [6]:

> Advanced Composites Group ZPREG o Resin stripes on one side of fabric

> Cytec Carboform o Resin impregnated random mat between the two fabric layers

> SP Systems SPRINT® (SP-Resin Infusion New Technology) o Resin between two fabric layers


3.2. Suitability LCM Processes for Mouldless Manufacturing The problems of vacuum leaks and foam core distortion under vacuum pressure experienced

during previously implemented mouldless VARTM method (Chapter 2) can be addressed in two

different ways. The first way is to use more robust bagging material and precise alignment

fixtures for manufacturing. The second way is to select and apply a process variant that readily

solves these problems.

The use of robust bagging material and precise alignment mechanisms will slightly increase the

cost of mouldless manufacturing. Though this is undesirable, the added cost will likely result in

better part quality. On the other hand, using an available technology usually comes with a

licensing fee. For some processes developed with the intention of reducing costs, this licensing

fee is usually small and is affordable by small or medium size companies. Thus, a process

variant is worth considering if it promises to solve the problems encountered in mouldless

VARTM.

VacFlo, Light RTM, RFI, and Semi-preg Infusion methods come with relatively high equipment

and material costs compared to conventional VARTM. Hence these processes are not

considered any further for mouldless manufacturing. Other processes such as VAP and Double

Bag VARTM promise better vacuum integrity as compared to conventional VARTM, but will

suffer similar core distortion under full vacuum pressure. Distortion of foam core under vacuum

pressure is a direct consequence of excessive 'pleats' (Figure 3.21) in the vacuum bag, which

conform onto the part unevenly when vacuum is applied to the part. Such pleats are

unavoidable when bagging complex parts with tough, disposable vacuum bags. These methods

will also have to rely on precise fixtures for controlling the part dimensions, which is somewhat


difficult to achieve when the part is surrounded with uneven vacuum bag pleats. This leads to

the conclusion that a vacuum bag that is form fitted to the shape of the part, will provide a

better solution to part shape retention in mouldless manufacturing. A form fitted vacuum bag

will allow application of uniform pressure around the entire tool, which may help minimize the

deflection of the foam core under vacuum pressure experienced in previously implemented

mouldless VARTM. CCBM is therefore a potential alternative to VARTM for mouldless

manufacturing. The competitive advantages of the CCBM process are listed in Table 3.3.

Table 3.3: Advantages and Disadvantages of CCBM process [24]

Advantages Reusable

Repairable

Many options for sealing the bags Translucent

Low cost mould requirements Less wastage

Material compatibility

Integrated manufacturing Workable

Robust

Suitable for all part sizes and geometry Better part quality

Description Tough and self cleaning bags that can last over 1000 cycles without any mould release application; improves the process repeatability. CCBM bags are easily repaired and restored to original condition when damaged by accident or rough handling. Unlike disposable vacuum bags, CCBM bags can be sealed in many different ways (i.e. flange seal, zipper seal etc.). Allows the operator to visually monitor the flow front and control the infusion. Existing moulds can be built up at low cost to make the mould for CCBM bag fabrication. CCBM reduces wastage of resin and bagging consumables compared to conventional VARTM. CCBM is compatible with almost all available matrix and reinforcement systems (i.e.: carbon-epoxy, fibreglass-polyester). CCBM allows manufacturing with integrated cores and ribs without the need for complex vacuum bag pleating. Minimal odour, no VOC's, easy storage and handling, and readily conforms to the shape of the mould. CCBM infusion offers improved vacuum integrity, reduced risk of bag leaks and eliminates the infusion complexities such as resin rich or resin starved areas caused by disposable bags. As the complexity and size of the part increases, CCBM makes vacuum bagging easier and cost efficient for low production quantities. CCBM can create cosmetically attractive bag side surface with no wrinkles.

Disadvantages

• Relatively high initial costs compared to traditional VARTM with disposable materials. • The durability of the bags is heavily dependent on the quality of bag manufacturing.


Figure 3.21: A part bagged with disposable vacuum bag [32]

CCBM bagging for the fuselage requires a multi-section bag, with a sealing mechanism

adjustable to tightly seal the bag around the entire part. An additional requirement is such that

the selected CCBM system and sealing method should be economical for part counts below 10,

in order to realize the economic advantages of mouldless manufacturing. From the review of

current CCBM techniques, it becomes apparent that this process is intended to be used for

manufacturing with rigid moulds, and hence can be expensive at the outset. Following the

detailed review of LCM methods, the CCBM process was selected for demonstration of

mouldless manufacturing; however, it was clear that significant process development and

assessment would be required to determine the most efficient and economical mouldless

CCBM setup.


CHAPTER 4. CCBM PROCESS DEVELOPMENT

This chapter presents process development work carried out to develop a suitable CCBM

technique for mouldless manufacturing. Various methods of bag-sealing and resin distribution

for CCBM process were identified. A series of manufacturing trials were conducted using flat

CCBM bags to determine the suitable techniques for mouldless manufacturing. Finally a Process

Value Analysis (PVA) was performed to evaluate the feasibility of CCBM for mouldless

manufacturing.

4.1. Mouldless CCBM manufacturing considerations

Applying CCBM in mouldless manufacturing setting requires a suitable bag sealing method that

does not depend upon rigid flanges. Additionally, efficient resin distribution techniques and

economic bag manufacturing strategies, if developed, would lead to further cost savings; hence

would make CCBM economical for mouldless manufacturing in low production quantities.

Several bag-sealing and resin distribution concepts for mouldless CCBM manufacturing,

investigated as part of this research are described in the following sections.

4.1.1. Sealing Mechanisms for Mouldless CCBM

Conventional CCBM process often uses extruded silicone sections, such as the Flexseal "D" to

create the vacuum seal. Some of the sealing configurations commonly used in the CCBM

process are shown in Figure 4.1. With the exception of interlocking seals, all seals require a rigid

mould surface and a flange as shown in Figure 3.13. This leaves the interlocking seal as the

most suitable sealing configuration for mouldless CCBM. However, this sealing method requires

the two parts of the seal to be positioned accurately on the top and bottom halves of the bag to

form an air-tight enclosure. This is somewhat difficult to attain considering the size and


geometric complexity of the fuselage. Hence an alternative, more economical sealing method

recommended in [34] was investigated. In this method, a tape with silicone based adhesive

backing (i.e. Airtech Flashbreaker® or Teflease®) is permanently bonded to the CCBM bag. The

taped surface of the bag is sealed against the tool using disposable sealant tape. This sealing

technique eliminates the high initial costs associated with the extruded silicone seals.

seal seal seal 16mm

l i I

Photo courtesy of Airtech Advanced Materials Group, Rubber Silicone Seals, [Online Catalogue], 2009, [Cited 12 Oct 2010], Available: http://catalogue.airtech.lu/product.php7product id=29&lane=EN

Figure 4.1: Common reusable seal configuration for CCBM

4.1.2. Resin Inlets and Outlets

One way to reduce the material waste associated with traditional VARTM is to use reusable

resin infusion lines. In the CCBM process, this is accomplished by permanently moulding the

resin infusion channels into the vacuum bag. To do this, two options were considered. The first

option was to mould the shape of tube into the bag by placing a waxed polymer tube on the

tool and fabricating the bag over it. The region of resin channels can be locally reinforced with

multiple layers of silicone to provide the necessary vacuum integrity. If this method was to

prove unsuccessful, a more conservative second option can be considered, in which extruded


http://catalogue.airtech.lu/product.php7product

silicone Omega Flow Lines (OF313) supplied by Airtech Corporation would be permanently

embedded into the CCBM bag during manufacture.

4.1.3. Resin Distribution in CCBM

The resin can be distributed across the part using either a disposable resin distribution medium

or by means of resin channels that are moulded into the CCBM bag (CIB infusion). The use of a

distribution medium is more conventional and has been proven to work in conventional VARTM

or CCBM settings [3,24]. On the other hand, CIB infusion is specifically advantageous for large,

complicated parts as it eliminates material and labour costs associated with the use of

disposable distribution medium. Additionally, the placement of such resin distribution channels

can be optimized to shorten resin infusion time and reduce resin consumption.

4.2. CCBM Manufacturing Trials

Three flat CCBM bags were manufactured to evaluate the functionality of various bag sealing

and resin distribution techniques discussed in section 4.1. The procedure employed in these

manufacturing trials and their outcomes are discussed in the following sections. Supplier

information of the materials used for this manufacturing trial is provided in Appendix A.

Arctek CCBM system [35] was chosen for these manufacturing trials, as it was the least

expensive, brushable CCBM system available. The required materials can be purchased in small

quantities. Important components of Arctek CCBM system are shown in Figure 4.2. It consists of

Proflex NS® silicone, which is supplied in 850 ml cartridges. This is a single-part, atmospheric-

moisture cure silicone. For 0.025 in. thick layer of Proflex NS®, the cure time is approximately 1

hour at 77° F/ 50% RH. An Arctek CCBM bag is fabricated by applying several layers of Proflex


NS® silicone over a rigid mould. A layer of Confortex® fabric is applied in-between the silicone

layers, which increases strength and tear resistance of the CCBM bag. Technical details and

manufacturing methods of this product can be found in [24].

Training DVD

Flexseal'D' reusable seal with adhesive backing

Proflex NS® silicone cartridges

Confortex® fabric

Figure 4.2: Components of Arctek CCBM system [35]

4.2.1. CCBM Manufacturing Trial # 1

In this trial, a 9 in. x 20 in. CCBM bag was manufactured for infusing a flat panel with the

assistance of disposable distribution medium. The bag featured Airtech Flashbreaker tape

bonded around the perimeter to facilitate sealing using disposable sealant tape. Half-circular

profiles were moulded into the bag to create resin inlet and outlet as shown in Figure 4.3. The

manufactured CCBM bag is shown in Figure 4.4. Step by step manufacturing method is provided

in Appendix B.


Flashbreaker tape

Application of first silicone layer

Half-circular profile to create resin inlet/outlet lines

Figure 4.3: CCBM bag manufacturing trial #1: tool setup

Top Surface

p

Resin Flashbreaker tape Inlet/Outlet

Flexible aluminum frame

Silicone tubes bonded to

reinforce the resin channels

{

Figure 4.4: CCBM bag manufacturing: trial #1 results

The results showed that using Airtac 2 spray adhesive to hold the Flashbreaker tape against the

tool (Appendix B) does not work for this sealing mechanism. Indeed, the spray adhesive

contaminated the surface of the tape, leading to permanent adhesion of the sealant tape to the

bag. Further, the half circular resin channels moulded into the bag collapsed under the applied

vacuum pressure. This bag was recovered by attaching a flexible aluminum frame around the

perimeter and bonding silicone tubes to locally reinforce the resin channels, as shown in Figure


4.4. This CCBM bag used Airtech Resinflow 60, disposable resin distribution media to speed-up

the infusion process.


A second CCBM bag was manufactured with slightly modified procedure to overcome the

difficulties in trial #1. In this trial, Airtech OF313 omega flow lines were embedded into the bag

to create the resin flow lines, and Teflease tape was bonded along the perimeter of the bag to

form the seal as illustrated in Figure 4.5. During manufacturing of the bag, Teflease tape was

held down to the tool surface with the use of a double-sided tape to prevent surface

contamination. The manufactured CCBM bag is shown in Figure 4.6. The problems encountered

in trial #1 were solved and much better quality bag (Figure 4.6) was produced in trial #2. Resin

infusion was facilitated with disposable resin distribution media. This bag is still in good

condition after 10 manufacturing cycles.

OF 313 Omega Flow line S> Resin Inlet/Outlet

Figure 4.5: CCBM bag manufacturing trial #2: tool setup

Teflease tape


I Top Surface

\ Teflease tape Resin Inlet/Outlet

OF 313 Omega Flow Line

Figure 4.6: CCBM bag manufacturing: trial #2 results


A third manufacturing trial was carried out to determine the feasibility of moulding resin

distribution channels into the bag. In this attempt, 0.125 in. diameter wire wax was placed onto

the flat tool using double-sided tape. Mould release wax was then applied on the tool surface

and CCBM bag section was manufactured over this setup, as shown in Figure 4.7.

.125 in. diameter „.„.;.:.•.;.««>:«:«:*:•«••.•::•:«::•:•:•

me wa* c

Tool preparation and

ig

Figure 4.7: Sample CCBM bag section with resin distribution channels


Following this trial, a complete bag section was manufactured to compare the infusion

characteristics and the part quality. Infusions with CIB method and disposable distribution

media are compared in Figure 4.8. In CIB infusion, the resin travels preferentially through the

distribution channels and then into the part, resulting in a flow front pattern as shown in Figure

4.8-(a). Resin infusion with disposable distribution medium results in a linear flow front as

shown in Figure 4.8-(b).

(a) (b)

Figure 4.8: Flow profiles of CIB method (a) and disposable distribution media (b)

4.3. Value Analysis of the CCBM process

CCBM manufacturing trials showed that with appropriate use of materials and low-cost bag

manufacturing techniques, CCBM can be made suitable for mouldless manufacturing of

complex geometry components. In order to determine the most economic CCBM setup for

mouldless manufacturing and assess its feasibility against mouldless VARTM method, a Process

Value Analysis (PVA) was conducted.


4.3.1. Introduction to Value Analysis

A product is considered valuable if it shows excellent performance characteristics and good

physical appearance, relative to its cost. In mathematical terms, the value of a product can be

expressed as [36].

_ (Performance + Capability) Cost { 4 1 }

_ Function Cost

The above expression shows that the value of a product can be increased either by minimizing

the cost or by maximizing the performance. In other words, the most valuable product will have

the highest functional worth (lowest cost to perform a given function). The concept of Value

Analysis was first developed in 1945, by Lawrence D. Miles, an engineer from General Electric

(GE) Company [37]. Mr. Miles found that with a systematic approach and clear understanding

of functional worth of the product, one could meet or improve product performance and

reduce its manufacturing costs. His approach to continuous improvement was called the Value

Analysis. The Value Analysis can be carried out on designs or processes, as an improvement

effort at any stage of a product life cycle. When several process variations exist, the Process

Value Analysis (PVA) is employed to determine the net value of each process variation. The

value of each process variation is then assessed to draw conclusions on the process feasibility.

Potential benefits of PVA include reduced material use and costs, reduced waste, reduced

distribution costs, improved profit margin, and improved customer satisfaction [36,37].


4.3.2. CCBM Process Value Analysis

The first step of the PVA was to identify the processes suitable for mouldless manufacturing

and develop a PVA matrix. Various sealing and resin infusion techniques investigated as part of

the manufacturing trials were assessed, to identify process variants applicable for mouldless

manufacturing (Table 4.1). In this PVA, it is assumed that each method described in Table 4.1

would yield similar part quality. Considering performance, the CCBM bag I, with extruded

silicone seal is easier to install during manufacturing than CCBM bag with sealant tape. This

benefit of the CCBM bag-l is realized at a slightly higher initial cost. In order to express the

functional worth as monetary values, all functional characteristics of the process variants were

grouped in terms of material and labour costs, with an initial labour rate of 20 $/hour.

Table 4.1: PVA matrix: processes for mouldless CCBM/VARTM

Process Variation CCBM bag 1 CCBM bag II CCBM bag III

VARTM

Description CCBM with extruded silicone seal and distribution medium CCBM with sealant tape and distribution medium CCBM with sealant tape and resin distribution channels embedded in the bag Traditional VARTM with disposable materials

Since CCBM bags are reusable for up to 1000 manufacturing cycles, the functional worth of

these processes were determined by estimating the cost per part for increasing part counts.

The total manufacturing cost per part was calculated by adding the fixed and variable costs

associated with the processes. All value assessment calculations were carried out using a

custom Microsoft Excel template created for this PVA. The assumptions made combined with

the equations used and the estimated values of each process are provided in Appendix C.


4.3.3. PVA Results

The total costs of manufacturing the fuselage for increasing part count were plotted at a labour

rate of $20/hour (Figure 4.9). The results show that the process development work, which

evolves to CCBM III, makes CCBM economical for 4 fuselages. This is an improvement from the

published manufacturer's data, which justifies the cost of CCBM for over 5 complex parts in a

conventional CCBM setting [35]. Thus, CCBM appears to be economical for mouldless

manufacturing. In order to assess the sensitivity of these results to labour cost, the cost

estimates were regenerated with a labour rate of $40/hr. The results shown in Figure 4.10,

suggests that doubling the labour rate increases the cost of the first part, but makes CCBM

economical for fewer parts (3 parts). This is primarily due to the significant labour cost savings

offered by CCBM for increasing part counts.

$3,000 -i

$2,500 - -

$2,000 •

o $1,500 — 4-* 0)

z

$1,000 -

$500

$0

0 1 2 3 4 5 6 7 Number of Parts

Figure 4.9: Cost of the fuselage for increasing part count at labour rate of 20$/hr


<po,uuu -

$2,500 -

$2,000 -

** ° $1,500-

z $1,000 -

$500-

$0-

-•-VARTM

— CCBM-I

— CCBM II

-*-CCBM III

3|r ^ " ^

— ~~ ~ z^^^^^^^^"^ — ~

— -

1 1 i 1 i 1

3 4

Number of Parts

Figure 4.10: Cost of the fuselage for increasing part count at labour rate of 40$/hr

4.3.4. PVA Conclusions

Mouldless CCBM is a viable option for manufacturing large complex geometry components.

CCBM III is cost effective after manufacturing 3 to 4 parts. Process robustness, repeatability and

part quality are likely to improve with proper implementation of mouldless CCBM.


CHAPTER 5. FUSELAGE MATERIAL SELECTION

This chapter provides a brief overview of the sandwich theory and discusses the choice of

materials for the new GeoSurv II fuselage.

5.1. Sandwich Theory

Sandwich constructions are widely used in many structural applications of advanced composite

materials. They consist of thin face sheets or skins adhesively bonded to both surfaces of a

relatively thick, low density core material. The core serves to increase the overall laminate

thickness, thereby keeping the Fibre Reinforced Polymer (FRP) skins apart. This leads to a

dramatic increase in flexural rigidity of the laminate for a small added weight. To illustrate the

sandwich principle, consider a symmetric sandwich beam in bending (Figure 5.1).

mffh

4 ¥

-» X

m/m

Figure 5.1: Sandwich beam subjected to three point bend

The stiffness coefficient (D) of this beam is given by [1]:

D, beam = {EI\ beam

= {El)core + {El\ skins

s 6 2 c 12

(5.1)


Where, E is the Young's Modulus of the material and / is the area moment of inertia of the

(d \ beam. For sandwich constructions featuring relatively thin skins — > 5.77 and relatively weak

\t )

core ' * • < > ! «

KE< C'

, the expression for D can be approximated as:

beam i

btd:

(5.2)

The above expression shows that the distance between the skins, "d" has greater influence on

the flexural rigidity of the beam compared to other variables in equation 5.2. The core

effectively increases this distance for a small increase in weight, thus making sandwich

construction stronger and stiffer than the corresponding monolithic counterparts. In bending of

a sandwich beam, the skin bears most of the bending stresses while the core predominantly

carries shear loads (Figure 5.2).

Top skin: In compression

VCore: In shear

^ -»-siV Neutral Axis

Bottom skin: In tension

Figure 5.2: Sandwich beam in bending

The core also serves to distribute the local loads to be carried by the skins over the entire

structure, without causing local failure; this makes sandwich construction an excellent design

solution for components exposed to impact and dynamic loading. The compression strength of


the core prevents local buckling of the skins and the shear strength coupled with good adhesion

between the skin and core aids in holding the skins in place under dynamic bending loads. The

latest core materials also offer excellent heat insulation, acoustic insulation, vibration damping

and fire resistance characteristics [2]. Detailed theory of sandwich structures can be found in

the literature e.g. [39-41].

5.2. Core Material for GeoSurv II Fuselage

The current fuselage uses Celfort® 300 EPS foam as the core material. This choice was primarily

driven by the lower cost of EPS foam and the proof of concept nature of the initial

manufacturing trials (Maley, [3]). Celfort® 300 has a nominal density of 1.42 lbs/ft3 and

compression strength of 30 psi. Initial infusions on test articles revealed that Celfort® 300

degraded due to surface resin absorption. Indeed, the infused sections distorted under full

vacuum pressure about 10 minutes after resin impregnation. Hence, during manufacturing of

the current fuselage, the foam parts were assembled and primed with a layer of fast cure epoxy

(West System 105) prior to infusion. The added resin layer allowed better handling of the core

during manufacturing and slightly increased its flexural stiffness and dimensional stability, at

the cost of increased weight. The manufactured component had up to 0.31 in. (7.8 mm)

deviation from the desired dimension, at several locations [3].

Replacing Celfort® 300 with an aerospace grade, structural core material promises potential

weight savings by eliminating the need to seal the foam surface prior to infusion. Core materials

specially formulated for structural applications and VARTM processing offer good mechanical

properties, lower surface resin absorption and could potentially improve the dimensional


tolerances of mouldless VARTM manufacturing. Hence, an investigation into state-of-the-art

structural core materials was performed to select a suitable core material for the new fuselage.

A wide range of foam, balsa wood and honeycomb core materials are available, with various

densities and finishes; all designed to meet specific structural and manufacturing requirements

of aerospace, marine, wind energy and transportation industries. The choice of core material

for the GeoSurv II fuselage must meet the requirements imposed by the nature of mouldless

VARTM manufacturing, listed in Table 5.1, in addition to complying with the UAS airframe

requirements outlined in Table 2.1. The following sections discuss various candidate structural

foam and balsa core materials that could be used for mouldless VARTM manufacturing.

Table 5.1: Manufacturing requirements for the core material

Requirements for Core Material

1. The core material shall be resistant to vacuum pressure (Compression Strength > 14.5 psi).

2. The core material shall feature nearly 100% closed-cell structure.

3. The core material shall be compatible with epoxy resins.

4. The surface finish of the core material shall lead to excellent skin-core adhesion.

5. The core material shall have excellent specific flexural stiffness to allow for near net shaped manufacturing.

6. The core material shall remain dimensionally stable after machining and during processing.

5.2.1. Structural Foam Core Materials

Structural foam cores are often preferred for Liquid Composite Moulding (LCM) processes. This

is mainly because most foam materials are easily machined or thermoformed. Their surface

finishes and cell packing densities can be tailored to meet a wide range of design requirements.


Also, most foam cores are readily available with optimized grooves and scrim-backings to

enable faster and better resin infusion.

Structural foams are processed from a variety of thermoset and thermoplastic polymers

including Polyvinyl Chloride (PVC), Polyurethane (PU), Styrene Acrylo-Nitrile (SAN), and

Polymethacrylimide (PMI), Polyethylene Terephthalate (PET), Polyester and Polyisocyanurate.

The mechanical properties, density and service temperature of structural foam cores can be

modified significantly by changing the ratios of chemical additives and various process

parameters such as pressure and temperature. The latest manufacturing technologies have

produced foam cores in densities ranging from 2 lbs/ft3 to 50 lbs/ft3, suitable for service

temperatures in the range of -184C to 260 C (-300° F to 500 F). They are available in sheets of

thicknesses up to 4 in. Some structural foam cores are also available in large block form to allow

for components to be machined as single integral bodies [42].

The selection process should carefully consider the mechanical properties, chemical resistance,

toxicity and costs of the available structural core materials. Table 5.2 describes some closed-

cell, semi-rigid foam cores that are potential candidates for the new fuselage. All of these

materials have better overall structural performance than Celfort 300 EPS foam utilized in the

current fuselage. They are available in various grades to meet the UAS certification standards

for fire, smoke and toxicity. The core selection for the new fuselage took into consideration, the

mechanical properties and costs of various structural foam materials, as described in section

5.2.4.


Table 5.2: Structural foam cores suitable for mouldless VARTM manufacturing'

Material Trade Names

Chemical Composition

Manufacturer

Unique Features

Closed-Cell Content

Density Range (lbs/ft3) Continuous thermal stability (° F) Grades

Cross-Linked PVC AIREX-C DIVINYCELL-H KLEGECELL-R Polymer based on cross linked Poly-Vinyl Chloride system • Alcan Composites • Diab Inc.

• Good peel strength • Compatible with most

resin systems • High thermal stability • Thermoformable above

212° F • Low water/resin

absorption • Contain gas under

pressure; possibility of out-gassing over time

> 95%

2 to 15.61

-328 to 158, 284 (High Temp.)

-» AIREX • C70: Universally

structural • C71: Elevated temp. • C52: Industrial

processing -» DIVINYCELL • H: High performance • HT: Aerospace grade • HP: Prepreg processing • F: low FST (Fire, Smoke

and Toxicity) • HCP : High density

Linear PVC AIREX-R

Polymer based on linear Poly-Vinyl Chloride system • Alcan Composites

• More elastic than cross linked PVC

• Good fatigue and impact resistance

• Thermoformable • Slightly lower

mechanical and thermal properties compared to cross linked PVC

>95%

3.75 to 8.70

131

• AIREX R63: Damage tolerant foam

PU LAST-A-FOAM NIDA FOAM PU

Polyurethane based chemical system

• General Plastics • Nida-Core Corp. • Resistant to most

chemicals and solvents

• More brittle and less fatigue resistant than PVCand SAN

• The surface at the resin-core interface tends to deteriorate over time

>95%

2 to 40

275, 320 (High Temp.)

• FR 6700: Aircraft grade

• FR7100: Modelling grade

• FR 10100: High temp.

• FR4300: Formable

• TR: Marine grade

SAN CORECELL

Styrene Acrylo-Nitrile based chemical system • SP Systems

(North America)

• Good machinability

• Resistance to water and fuel oil

• Minimal density variation

• No out-gassing problems

• Compatible with most resin types

• Thermoformable

>95%

3.6 to 19.7

185, 230 High Temp.)

• Core-Cell A: For dynamically loaded structures

• Core-Cell P: Prepreg processing

• Core-Cell T: For decks and interiors

• Core-CellS: sub-sea applications

2 Information provided in this table were obtained from the manufacturers' data. The references are provided in Appendix D.

3 Problem areas are identified in bold and italicized font.


Table 5.2 Continued...

Material Trade Names

Chemical Composition

Manufacturer

Unique Features

Closed-Cell Content

Density Range (lbs/ft3) Continuous thermal stability (° F) Grades

PMI ROHACELL

Polymethacrylimide based chemical system

• Evonik Industries /Degussa

• Resistant to organic solvents such as benzene, xylene and monostyrene

• Resistant to fuel constituents and solvents for paints

• Features better mechanical properties compared to other structural foam cores

• Optimized for LCM processes

• Relatively lower resin absorption than other foams

2.0 to 6.9

320

->Rohacell • RIST: Low surface resin

absorption • RIMA: Finest cells for

minimum surface resin absorption

• IG: Industrial grade • A: Aircraft grade • WF: Heat resistance

grade • XT: Extended temp. • S: Easy to shape and

machine

PET AIREX-T NIDA FOAM PET Poly Ethylene Terephthalate based chemical system • Alcan Composites • Nida-Core Corp.

• Good high temp, stability

• Easily machined and thermoformed

• Chemically stable

> 95%

6.3 to 20

212, 302-392 (High Temp.)

• AIREX-T 90: Easy processing

• AIREX-T 91: Easy processing

• Nida Foam PET100/150: Structural

Polyester AIRCELL

Polymerized cross-linked aromatic polyester system

• Polyumac Inc.

• Good impact and fatigue resistance

• Durable and resilient

• flame retardant, non-friable

Closed content cell comparable to PVC

4 to 36

-320 to 165

• Aircell T: structural

Polyisocyanurate ELFOAM

Polyisocyanurate based chemical system • Elliott

Company

• Excellent chemical resistance and resin compatibility

• Easily machined, perforated and cut

• Class 1 flammability rating

• Good thermal insulation

Closed-cell content comparable to PU 2 to 6

-297 to 298

• Elfoam P: Structural series


5.2.2. Balsa Wood Cores

Balsa wood is another core material known for its high specific compression strength and

stiffness. It is dried and prepared from naturally harvested balsa lumber in the end-grain

configuration, as shown in Figure 5.3. In the end-grain configuration, balsa core features lean

closed cells tightly packed and oriented perpendicular to the plane, forming a closed-cell,

honeycomb like structure at microscopic level. This leads to its ability to resist high compression

and dynamic loads. Apart from this, balsa cores are fire resistant and act as thermal and

acoustic insulators in sandwich constructions [42,44].

Figure 5.3: Balsa wood core-end grain configuration [43]

The main disadvantage of balsa core is its high density, with lowest minimum values ranging

between 5.5 lbs/ft3-6 lbs/ft3. The density factor is further aggravated with balsa's high surface

resin absorption characteristics, making it non-preferable for weight critical applications. The

three main manufacturers of advanced structural core materials, Diab Corporation, Alcan

Composites and Nida Core Corporation are competing to produce lightweight-consistent


density balsa cores by utilizing advanced processing techniques. Balsa cores are now pre-sealed

with primers specifically formulated for lower surface resin absorption, and are available in

sheets of thicknesses up to 4 in. or as contoured blocks held together with fibreglass scrim [44].

5.2.3. Other Core Materials

Other core materials that may be utilized in VARTM applications include the following:

> Cedar wood: It is a natural lumber often used as core material in strip-plank

construction. It features grains running parallel to the plane, offering some stand

alone bending rigidity compared to balsa wood. However, it exhibits poor impact

resistance, low torsional rigidity and compression strength compared to balsa wood

[44].

> Airex PXc®: It is glass fibre reinforced PU foam supplied by Alcan Composites. The

foam exhibits exceptional dimensional stability, chemical and thermal resistance. This

material is available on custom order and hence is very expensive. Also, the minimum

available density of this material is higher than 6 lbs/ft3 density range [45].

> Airex PXw®: It is continuous glass fibre fabric reinforced PU foam supplied by Alcan

composites. The foam is uniquely formulated to exhibit good flexural rigidity on its

own, allowing it to be used with or without face sheets. Airex PXw® also offers

exceptional dimensional stability, chemical and thermal resistance. This material is

available on custom orders and therefore is very expensive. Further, its minimum

available density is larger than 6 lbs/ft3 range [46].


> Foam Filled Honeycomb: Nida Core Corporation and MGI Inc. supply PU foam filled

honeycomb core materials at densities ranging from 5 lbs/ft3 to 20 lbs/ft3. They

combine the properties of honeycomb and foam materials to offer exceptional

compression and shear performance. It is a closed-cell core that can be used with

VARTM and other closed moulding processes. Sealing the honeycomb with foam

imparts increased stand-alone flexural rigidity to this material, which makes it

preferable for mouldless VARTM application. The major drawback is that

manufacturing of PU filled honeycomb is an emerging technology; hence the flexural

rigidity has not yet been quantified. The mechanical properties of 5.56 lbs/ft3 density

material from MGI Inc. were found to be lower than the cross linked PVC cores at the

same density. Also, its machinability, workability and resin absorption characteristics

need to be characterized prior to using them for VARTM applications [47,48].

5.2.4. Core Selection for GeoSurv II Fuselage

The mechanical properties and costs of the structural core materials were considered in order

to select an appropriate core material for mouldless manufacturing. Complete profile of

mechanical properties attributed to various foam and balsa core materials, along with the

supplier details are given in Appendix D. Balsa cores, though exhibiting mechanical properties

superior to foam cores, were not considered in the selection process, as they are too heavy in

their lowest available density. Foam filled honeycombs Airex PXc® and PXw® were also

excluded from the selection process due to limited supply and high costs. This narrowed the list

down to structural foam cores as the primary candidates for the new fuselage.


The structural foams listed in Table 5.2 were assessed for their performance in shear and

compression. This assessment identified PMI (Rohacell), crossed linked PVC (Airex C, Divinycell),

and SAN (Corecell A) as the top three candidates for the new fuselage. Their properties relative

to cost were compared to make the final choice. Since not all of the materials are available at

the same densities, the mechanical properties of the top three foam cores were first

normalized at 4 lbs/ft3 using linear relations. The normalized properties of the top three foam

cores along with their nominal costs are given in Table 5.3. These findings are compared in

Figure 5.4.

Table 5.3: Normalized mechanical properties of the most structural foam cores

Material (4 lbs/ft3)

Rohacell Airex C Corecell A

Compression Strength

(psi)

189 144 56

Percent difference

compared to Corecell A (%)

238 157 0

Shear Strength

(psi)

157 132 89

Percent difference


77 48 0

Cost of 0.5 in thick

material ($/ft2)

11.35 2.50 4.15

Percent difference


173.58 -39.76

0.00

Rohacell (PMI) Airex C (PVC) Corecell (SAN)

a Compression Strength • Shear Strength a Cost

Figure 5.4: Comparison of the selected foam materials at 4 lbs/ft density


The results demonstrate that cross-linked PVC (Airex C) foam cores offer good mechanical

properties at a reasonable cost. Though the latest PMI (Rohacell) foams exhibit 20% better

overall structural performance, their costs are significantly higher than the PVC cores. This

would be a definite concern for the low-cost aspect of this process development work. Hence,

cross-linked PVC foam was chosen as the core material for the new fuselage. Alcan Airex C 70

series structural PVC foam at various density grades, donated by Alcan Composites Inc. was

used throughout this research.

5.3. Selection of Matrix and Reinforcement Materials

The current fuselage (Maley, [3]) was manufactured with SC-780 toughened epoxy matrix

supplied by Applied Poleramic Inc. (API) and AGP-370-5H satin carbon fibre fabric

reinforcement supplied by Hexcel Corporation. The trade studies for the choice of these

materials were carried out as part of the previous mouldless VARTM research and can be found

in [3]. Preliminary Finite Element Analysis (FEA) carried out in [49] showed that most of the

current fuselage structure was overdesigned. In order to reduce the structural weight, Style#

94132-4H satin carbon fibre fabric supplied by BGF industries was selected for the new

fuselage. Additionally, BGF Style# 106-Plain E-glass fabric was chosen as the finishing layer to

provide a smooth surface finish on the outside surfaces of the fuselage. The specifications of

the current and the newly selected fabrics are compared in Table 5.4.

Table 5.4: Carbon fibre fabric specifications

Specification Supplier Fibre Type Tow Weave Style Weight

Hexcel-AGP 370 5H Satin Hexcel Corporation AS4 (Medium Modulus) 6K 5H Satin 11.1 oz/yd2

BGF-94132-4H Satin BGF Industries T300 (Medium Modulus) 4K 4H satin 5.8 oz/yd2

BGF-106-Plain BGF Industries E-glass IK Plain 0.72 oz/yd2


In the new fuselage, PR 2712 infusion epoxy from PTM&W Industries Inc. was substituted for

SC-780 epoxy. Both resin systems are comparable in terms of infusion characteristics and

mechanical properties. The choice was primarily attributed to the lower cost and supply of PR

2712 epoxy from local distributor Composites Canada. The material supplier details for the

matrix and reinforcement materials are given in Appendix A.

5.4. Material Selection for Rigid Inserts

The GeoSurv II fuselage has several structural joints (i.e. bolted pin-joints) through which

discrete loads are introduced into the sandwich structure. Due to the relatively low strength of

the foam core, such locations are susceptible to local failure and must be reinforced with rigid

inserts. The current fuselage uses Delrin® inserts, which were bonded using a structural epoxy

adhesive. Delrin® is machinable and exhibits high specific strength. However, it does not adhere

well to epoxy, often causing failure in the bondline where the insert interfaces with the

fuselage. Hence, alternatives including Fibreglass Reinforced Polymer (FRP), Poly-Ether Ether

Ketone (PEEK™) and glass filled PEEK inserts were considered for the new fuselage. Costs and

mechanical properties of the aforementioned materials are shown in Figure 5.5.


t

500000

450000

400000

350000

300000

250000

200000

150000

100000

50000

0

• Specific Compressive Strength

• Unit Cost

140

120

100

80

60

40

c 3

Fibreglass Reinforced Polymer PEEK 30% GLASS FIUED PEEK DELRIN

Figure 5.5: Comparison of inserts for sandwich assembly

PEEK and Glass Filled PEEK are both considerably lighter than FRP and meet the strength

requirements of the fuselage. However, they are relatively expensive, difficult to machine and

exhibit poor impact characteristics. FRP inserts are relatively inexpensive compared to PEEK and

exhibit good impact properties, specific strength and machinability. Hence, FRP inserts were

selected for the new fuselage. The specifications and supplier information of the inserts are

included in Appendix D.

Note: the unit cost was estimated for 1 in. diameter rods.


CHAPTER 6. FUSELAGE REDESIGN

The redesign objectives and design changes made to the GeoSurv II fuselage are described in

this chapter. An improved GeoSurv II fuselage model is also presented.

6.1. Redesign objectives

The main objective of the fuselage redesign is to improve upon the current design with the

intention of producing a near-net-shape fuselage by mouldless manufacturing. The new

fuselage shall replace the current fuselage in the GeoSurv II prototype, while interfacing with

already manufactured sub-assemblies, such as the wing. The redesign goals and limitations

established based on these intentions are summarized in Table 6.1.

Table 6.1: GeoSurv II fuselage redesign: goals and limitations

Redesign Goals > Reduce weight and improve part quality > Improve the manufacturability of the fuselage

(DFM) > Reduce the steps required to finish the

fuselage: Integrated Manufacturing (IM) > Improve dimensional tolerances (near-net-

shape manufacturing)

Redesign Limitations > Design for mouldless manufacturing: foam

core sandwich construction > The new fuselage shall replace the current

fuselage in the GeoSurv II prototype: No major changes to the Outer Mould Line (OML)

The fuselage redesign work-flow diagram is shown in Figure 6.1. DFM and IM principles were

applied to improve the manufacturability of the current fuselage by mouldless CCBM. The

underlying principle behind this work is continuous product and quality improvement, while

ensuring that the proposed design specifications are in fact achievable with the selected

process. In this work, several DFM and IM principles outlined in [50] were considered, including

minimum part count, ease of fabrication and assembly. The following section presents the

design changes and the improved design concept of the new fuselage.


Problem Definition Review of the current fuselage

design/manufacturing (Chapter 2)

Selection of New Processing Method and Materials

(Chapters 3-5)

New Fuselage Design (Chapter 6)

Figure 6.1: Fuselage redesign work-flow diagram

6.2. Design Changes to the GeoSurv II Fuselage

The new fuselage walls were designed with edge-stiffened panels in place of the current

'tapered' panels, as shown in Figure 6.2. The required flexural rigidity can be achieved with

thinner edge-stiffened panels than with tapered panels (validated through FEA), resulting in

potential weight reduction and increased internal volume of the fuselage. Additionally, the use

of edge-stiffened panels increases the lengthwise stiffness of the fuselage walls and might

potentially reduce the part distortion experienced during mouldless manufacturing.


Current Fuselage New Fuselage

Figure 6.2: Current and the new fuselage wall design (units: in.)

In the new design, rigid inserts will be embedded into the foam core prior to fabric layup and

infusion. A bolted sandwich joint that results from this approach is compared to a current

fuselage joint in Figure 6.3. In its current configuration, rigid inserts are secondarily bonded into

the fuselage and washers are sandwiched in-between the nut and bolt. This approach requires

oversized washers to properly transfer out-of-plane loads into the fuselage. The reason is to

minimize stresses within the secondary bond line at the outer surface of the insert. In the new

joint design, the loads are transferred into the fuselage through normal contact of the bolt with

the skin and the insert. Additionally, the bonded area at the skin-insert interface also aids in

distributing the loads into the structure, thereby eliminating the need for oversized washers.

The use of small washers combined with in-situ bonding of skin and inserts is expected to result

in a lighter yet structurally sound bolted joints.


Bushing [Optional)

Current sandwich bolted New sandwich bolted assembly assembly

Figure 6.3: Current and the new bolted sandwich assembly

The new fuselage also incorporates a landing gear attachment plate to facilitate mounting of

the main landing gear, as shown in Figure 6.4. The new design shifts the attachment point of

the landing gear 5.5 in. aft of the current design. This design change simplifies the complex

swept-back main landing gear design into a straight-beam configuration, as shown in Figure 6.5.

Eliminating the sweep from the landing gear also reduces the moment experienced during one

wheel landing from 12500 in-lbs to 9200 in-lbs. Thus, this design change offers potential weight

savings in the main landing gear strut and the landing gear attachment panel.


t ^ o )

* \ Plate placement

Landing Gear Attachment Plate

Figure 6.4: New landing gear attachment plate

WL

<—> 6.6 Inches

<-> 3 Inches

Current Landing Gear Configuration New Landing Gear Configuration

Figure 6.5: Current and the new landing gear configurations

Several other design changes were made to improve the manufacturability and minimize the

effort required to prepare the fuselage for assembly. These include: extension of the straight

section of the fuselage wall (Figure 6.6), geometry modification to facilitate mounting of the

shear-pins (Figure 6.7), increased core thickness at the location of the fasteners (Figure 6.8), an

extension on the front bulkhead to substitute for the secondary bonded nosecone bridge


(Figure 6.9) and reinforcement for mounting of the flight avionics rack (Figure 6.10). In the

process of applying Integrated Manufacturing (IM) principles, the bay separator panel, shown in

Figure 2.5 was not included in the foam components, as it would complicate the vacuum

bagging process. The redesigned fuselage model is shown in Figure 6.11.

m^%^mm


Figure 6.6: Fuselage wall straight section extension

Current Design

New Design

Figure 6.7: Design modification at the fairings



Figure 6.8: Increased core thickness at the locations of the fasteners


Figure 6.9: Core extension to mount the nosecone

High density (12.5 lbs/ft*) PVC foam reinforcement for

mounting avionics rack

Figure 6.10: Reinforcement for flight avionics rack

CHAPTER 6. FUSELAGE REDESlGf

Current Design New Design

Figure 6.11: Current and the Modified fuselage concept models


CHAPTER 7. DESIGN OPTIMIZATION

This chapter presents the Finite Element Analysis (FEA) carried out to optimize the new fuselage

and the experimental study conducted to verify selected FEA results.

7.1. GeoSurv II Fuselage FEA

A Finite Element Analysis (FEA) was carried out on the GeoSurv II fuselage, to simulate the

structural performance of the new design under flight manoeuvre and landing loads. The

objective of this FEA was to obtain complete stress-strain profiles at critically loaded areas of

the fuselage, to allow for failure predictions and layup optimization of the new design. The FEA

was carried out using Abaqus 6.8-2, as a linear-elastic stress analysis with several other

simplifying assumptions (discussed in the following sections). The FEA results for selected

loading conditions were verified experimentally using test coupons. The following sections

describe the FEA procedure and results.

7.1.1. FE Model Construction

The parts required to create one half of the fuselage assembly were modelled using the Abaqus

CAE modeller. The parts list includes foam core, skin, wing carry-through spar, bushing, engine

pins, nose and main landing gear pins and wing shear pins. The foam core was modelled as one

3-D deformable solid part and the skin was created from the core, using the "create shell: from

solid" feature available in Abaqus CAE modeller. The carry-through spar was modelled as a

circular shell extrusion and all other components were created as 3-D deformable solids with

appropriate features, as shown in Figure 7.1.


The rationale behind the half model was to impose the symmetry condition in the FEA and

hence, reduce the computational time required to run the simulations. To further reduce the

simulation times and expedite the meshing process, several simplifications were made to the

part geometries. The geometric simplifications were as follows:

> Pins were modelled as circular extrusions.

> Fillets and chamfers farther away from the highly loaded areas were excluded from the

part geometry.

> Skins at the top and bottom edges of the fuselage that do not contribute to its load

carrying capabilities were not modelled.

Figure 7.1: Parts modelled for the fuselage FEA


7.1.2. Material Properties

The following sections describe the material properties and assumptions employed for various

parts of the fuselage assembly.

Foam Core:

Airex C structural PVC foam at different density grades, supplied by Alcan Composites was used

as the core material for the new fuselage. Previous research in [51] has shown that PVC foams

are transversely isotropic. However, the variation of material properties along different

directions, which is caused by the differences in the cellular development during the foam

expansion process, is rather small and hence ignored for the purpose of this FEA. Thus, the

conditions: Ei=E2=E3 and Gu= G13= G23 were established in the FEA. The values of E and G were

obtained from the material properties published by the manufacturer.

There was no measured or published value of Poisson's ratio (u) for PVC foams. This was

because under uniaxial loading, PVC foams show nearly no deformation along the transverse

axis of the test direction and the respective strains are rather difficult to measure. Additionally,

the PVC foams do not exactly obey the linear-elastic relation between E, G and u. For this

reason, a Poisson's ratio of 0.32 was assumed [52] along with the condition U\i = U13 = 023-

Various density grade PVC foams were modelled with properties as shown in Table 7.1.

Material sections having properties of various density PVC foams were established in the FEA

using the "Engineering Constants" entry, found in the property module of Abaqus CAE.

Subsequently, the parts were assigned appropriate material orientations using local coordinate

systems.


The foam core was created as one solid component and was partitioned into different sections.

These partitions were assigned different section properties to represent the parts made with

different density PVC foams (Table 7.1). Thus, a multi-part fuselage foam assembly was created

as single part in the FE model. This section property formulation assumes a perfect bond

between different foam parts, which is a reasonable assumption, given that the adhesive used

to bond the foam parts is stronger than the foam.

Table 7.1: Properties of Airex C PVC foam

Airex C Structural PVC Foam

C70.40

C70. 55

C70. 75

C70.90

C70. 200

Density (lbs/ft3)

2.5

3.7

5.0

6.2

12.5

Ei=E2=E3

(psi) 5947

10000

15080

18850

40600

U l 2 = Ul3 = U23

0.32

0.32

0.32

0.32

0.32

Gl2= Gw= G23 (psi) 1900

3190

4350

5802

10900

Skin:

The new fuselage features carbon-epoxy composite skins with BGF-Style# 94132-4 H satin

carbon fibre fabrics and PTM&W-PR 2712 epoxy. In the FEA, each ply of the woven fabric was

modelled as two unidirectional laminae, with half the total ply thickness. Since no

characterization data were available for these materials at the time of this analysis, the lamina

properties were derived from in-house coupon test data (Table 7.2) available for the current

fuselage materials. The current fuselage skin features Hexcel-AGP-370 5H satin carbon fibre

fabric and API-SC-780 epoxy. Both SC-780 and PR 2712 resin systems have similar mechanical

properties and are compatible with carbon fibre fabrics. The major difference between Style#


94132 and AGP 370 fabrics lies in the fibre type and the crimp angle. The AGP 370 fabric

features AS4C type fibres with 6k tow count while the Style# 94132 fabric features T300 fibres

with 3k tows. However, both fabrics are compatible with epoxy resins, which allows for realistic

prediction of lamina properties for the new fabric-resin combination, from the test data

available for the previous reinforcement and matrix materials. The derivation of the new

lamina properties used in the FEA is described in Table 7.3.

Table 7.2: Properties of the current fuselage materials

Material

AGP370 5H satin carbon fibre fabric /SC-780 epoxy

Tensile Strength

(psi)

122400

Tensile Modulus

(psi)

9325900

Strain to Failure

(Tension)

0.013

Shear Strength

(psi)

8600

Shear Modulus

(psi)

102250

Strain to Failure

(In-plane shear)

0.05


Table 7.3: Derivation of lamina properties for the new fuselage

Available/Estimated Properties

E 1 = 9325900 psi

E2 = 932590

G12 = 102250

G13 = 10225

G23 = 10225

u = 0.3

Tensile strength = 122400 psi

Strain to failure (Tension) = 0.013

Shear strength = 8600 psi Strain to failure (In-plane shear) = 0.05

Source

In-house coupon testing Hart Smith's 10 % rule [53] In-house coupon testing Hart Smith's 10 % rule [53] Hart Smith's 10 % rule [53] Typical for carbon-epoxy composites In-house coupon testing

In-house coupon testing

In-house coupon testing In-house coupon testing

Knock Down Factor and Rationale

N/A: T 300 and AS4C fibres have similar moduli N/A

N/A: In-plane shear is a matrix dominated property N/A

N/A

N/A

-10% (T300 vs. AS4) +10% (Crimp angle: 3k vs. 6k) -10% (Woven fabric = Unidirectional lamina assumption in the FEA) [54] +10% (T300 vs. AS4) -20% (Woven fabric = Unidirectional lamina assumption in the FEA and additional safety factor of 10%) [54]

15% of the tensile strength; based on in-house test data [3] -20% (Woven fabric = Unidirectional lamina assumption in the FEA and additional safety factor of 10%) [54]

Derived properties

E 1 = 9325900 psi

E2 = 932590 psi

Gi2 = 102250 psi

G13 = 10225 psi

G23 = 10225 psi

u = 0.3

Tensile strength = 101160 psi

Strain to failure (Tension) = 0.011

Shear strength = 15174 psi Strain to failure (In-plane shear) = 0.04

Similar to the foam core, the skin was modelled as a single piece conventional shell in Abaqus

and was partitioned into small sections to facilitate section property assignments and local

mesh refinements. Highly loaded areas of the skin were locally reinforced by assigning different

composite layups.

Carry-through spar, inserts, bushings and pins:

The carry-through spar was assigned conventional shell composite properties (Table 7.2) with

layup [0°/90°]g. The steel pins, Delrin® bushings and fibreglass inserts were modelled as


isotropic materials with properties given in Table 7.4. Since the pins and bushings were

employed as connecting elements, to distribute the loads into the structure, an isotropic

property formulation was sufficient for this FEA. The inserts were included in the FE model as

partitioned sections of the foam core, while the pins and bushings were modelled as separate

parts and were assigned their respective material properties. The FEA representation of the

inserts assumes a bondline stronger than the foam, at the insert-foam interface. This

assumption is acceptable, when a structural epoxy adhesive is used for bonding the inserts into

the foam core. Additionally, the inserts are much stronger than the foam and are not expected

to fail before the core or the skin. Hence, to simplify the FEA process, they were also modelled

with isotropic element properties.

Table 7.4: Properties of pins and bushings

Part

Pins Bushings Inserts

Material

Stainless Steel (SS) 316 Delrin®

Fibreglass Rod

Young's Modulus (psi)

27992283 430000 2800000

Poisson's Ratio(u)

0.3 0.3 0.3

7.1.3. Part Meshing Considerations

The foam core, inserts, bushing and pins were meshed with C3D8R elements, which are 8-node

linear bricks with reduced integration and hourglass control. This is a solid, 3-D stress element

in Abaqus that is computationally effective for simplified linear-elastic analysis. The composite

skin around the foam core represents a typical plane-stress scenario, in which the thickness of

the skin is significantly smaller than the other in-plane dimensions. Hence, they were modelled

with S4R elements, which are 4-node, quadrilateral conventional shell elements, featuring

reduced integration, hourglass control and finite membrane strains. It is a common practice to


model plane stress states in this fashion due to improved computational efficiency and lack of

through-thickness effects. With the conventional shell formulation, the skin geometry was

modelled as a shell and its thickness was defined by specifying a value and a stacking direction

in the property module of Abaqus. Similar to the skin, the carry-through spar, which features 8

plies of carbon-epoxy composite, was also modelled as a conventional shell using S4R elements.

The fuselage main frame was meshed by creating a series of partitions to sub-divide the

geometry into simple and compatible sections. These partitions (Figure 7.2) facilitate the mesh

propagation in a uniform manner through the entire fuselage geometry. Partitions were

required in the regions where geometric changes or interactions occur. Additionally, the pin

holes were partitioned to be contained in simple square sections, having partitions in a radial

fashion towards the holes. This partitioning strategy, illustrated in Figure 7.3, allows biasing the

mesh seeds and locally refining the mesh near the highly stressed pin-holes. Thus, the areas

away from the critically stressed regions can be meshed in a coarser yet uniform manner to

improve the computational efficiency. All other components of the fuselage assembly were

relatively simple geometries and were meshed by creating partitions, in a similar manner.


Figure 7.2: Partitions created on the fuselage skin for meshing

Figure 7.3: Mesh refinement near the pin holes


7.1.4. FE Model Assembly and Constraints

The assembly of the fuselage model was created by importing the required part instances into

the Assembly module of Abaqus and applying appropriate position constraints. The final FE

model assembly of the fuselage is shown in Figure 7.4. Following the assembly process, tie

constraints were established at the regions where the pins and the bushings come in contact

with the fuselage structure. In a more realistic attempt to represent bolted pin joints, an analyst

would introduce a contact algorithm in the analysis. This approach is more complex and beyond

the scope of this FEA. In order to keep the tie constraints formulation realistic for bolted joints,

the surface to be tied were carefully chosen based on the loading conditions. Thus, for loading

along a particular direction, as in the case of shear pins, only one half of the pin surface was

tied to the fuselage structure. For all other pins subjected to simultaneous loads in different

directions, the entire pin surface was tied to the fuselage structure, as shown in Figure 7.5.

Figure 7.4: Fuselage FE model assembly


Figure 7.5: Tie constraints established between the pins and the fuselage structure

7.1.5. Analysis Steps, Loads and Boundary Conditions

The FEA was implemented in two general static analysis steps: an in-flight step and a landing

step. During the in-flight step, the loads from the wings, engine and payload were applied at

the maximum design ultimate load (DUL) factor of 10.1, which stems from the maximum

positive design limit load (DLL) factor of 6.75 multiplied by the safety factor of 1.5 [55,56].

During the landing step, all in-flight loads were reduced to the maximum landing load factor of

3.5 and the landing loads were introduced. During both analysis steps, the base plate and the

top of surface of the fuselage were fixed and symmetry conditions were applied across the

surface as shown in Figure 7.6. The load cases considered are illustrated in Figure 7.7 to Figure

7.9. The load factors and the loads were obtained from the GeoSurv II design report registry

references [55-57].


BC: Fixed- top skin and core

BC: XSYMM: Symmetry about the X= Const, plane

Figure 7.6: Boundary Conditions

Engine Loads:

->Torque at 0.8 RPM 921 in lb/bolt

->Weight 239 lbs 59.75 lbs/bolt

Lift Moment: 8056 in. lbs applied as shear forces at the shear pins

Lift load: 1912 lbs total, applied at the tip of the carry through spar

Figure 7.7: Wing lift and engine loads


Main Landing Gear Loads: -^Vertical Landing load: 400 lbs or 100 lbs/in. applied on the pins -^•Side Moment: 9200 in. lbs applied on the area of the washer -^Vertical landing load 600 lbs applied as the pressure load on the area where the landing gear contacts the attachment bracket

Figure 7.8: Main landing gear loads

Air Data boom loads: Weight: 26 lbs Moment: 291 in. lbs

Nose Landing gear loads: Vertical Load: 90 lbs Drag: 54 lbs Side load: 56 lbs

Figure 7.9: Mission avionics and nose landing gear loads


7.1.6. Mesh Independence of the Results

A parametric study was performed on the substructure model (Figure 7.10), to determine the

density of the mesh at which the results converge. Throughout the study, the skin and the core

were refined at the same rate, starting from a coarser mesh (1880 total elements), as shown in

Table 7.5. The mesh refinement on the skin is illustrated in Figure 7.11. This study revealed that

the mesh independence of the results with 5% convergence on the Von-Mises stress and 0.2 %

convergence on the displacement occurs at element size of 0.09 in. with the curvature factor of

0.003 in. The corresponding mesh convergence graph is shown in Figure 7.12.

BC: Fixed

Load: Pressure

BC: Symmetry

Figure 7.10: Substructure model used for parametric study


Level 1 Level 11

Figure 7.11: Mesh refinement: parametric study level 1 to level 12

Table 7.5: Parametric Study Results

Level

1 2 3 4 5 6 7 8 9 10 11

Element Size/Curvature

control

0.5/0.1 0.25/0.05 0.2/0.01 0.15/0.01

0.125/0.01 0.11/0.01 0.1/0.005 0.1/0.003

0.09/0.003 0.0875/0.001 0.085/0.001

Number of Elements

(n)

1880 5994 13564 23058 36421 52879 72459 85764 111886 146282 154510

Nodal Displacement

(Us)

0.0716217 0.0776539 0.0786073 0.0787259 0.0785697 0.0786482 0.0785664 0.0785424 0.0786624 0.0786709 0.0786694

% Difference

8.08% 1.22% 0.15% 0.20% 0.10% 0.10% 0.03% 0.15% 0.01% 0.00%

Max Von Mises Stress (psi)

8350 9774 7970 7946 11220 9549 10510 10690 10320 10320 10310

% Difference

15.71% 20.33% 0.30% 34.16% 16.09% 9.58% 1.70% 3.52% 0.00% 0.10%


11500

11000

10500

=10000

g 8500

8000

7500

7000

+ 5%

20000 40000 60000 80000 100000 120000 140000 160000

Number of Elements (N)

Figure 7.12: Von Mises stress convergence (5%)

Following the convergence study the fuselage was meshed in accordance to the partitioning

and meshing strategies discussed in section 7.1.3. Highly loaded areas were meshed with fine

elements (size 0.09 in. or smaller) and all other regions were meshed in a slightly coarser yet

uniform fashion to improve the computational efficiency. Upon completion, meshes were

verified to correct the elements with poor aspect ratios and internal angles.

7.1.7. FEA Simulations

A series of FEA simulations was executed in order to determine the optimum fuselage layup.

Starting from a base single ply layup, the simulations were iterated with modified layups until

an optimum layup for the fuselage was found. After each simulation, the Von Mises stress, Si2


and 13 results were compared against the available material properties, based on which the

layups were modified for the next simulation run.

The final FE model (Figure 7.13) was constructed with 1,213,298 structural elements and

required a minimum of 16 GB computer memory to run the simulation. These simulations were

executed using computer clusters featuring quad-core processors and maximum memory

capacity of 32 GB. At this configuration, each simulation run took approximately 9 hours to

completion.

Figure 7.13: Fuselage FE model


7.1.8. FEA Results

The optimized structural foam core components of the new fuselage are shown in Figure 7.14.

Medium to high density PVC foam (5 lbs/ft3 -12.5 lbs/ft3) is used at highly loaded areas such as

the front bulkhead, rear bulkhead, rear walls and the landing gear attachment. The base plate

and the front walls that do not experience high flight manoeuvre or landing loads were

constructed with low density, 3.7 lbs/ft3 PVC foam. The locations of bolted pin-joints were

reinforced with FRP inserts to prevent core crushing under discrete loads.

r Z l 3.7 lbs/ft3

Figure 7.14: Optimized foam core for the new fuselage


I I

I

lply[±45°] t= 0.012 in.

2 plies [±45°, 0°/90°] t= 0.024 in.

3 plies [+45°, (0o/90o)2] t= 0.036 in.

4plies[±45°,(0790)2, ±45°,] t= 0.048 in.

8 plies [(±45°)4, (0790°) t= 0.096 in.

Figure 7.15: Optimized skin layup for the new fuselage

Shown in Figure 7.15 is the optimized skin layup for the new fuselage. The locations of pin-

joints and carry-through spar were reinforced with either three or four plies of the carbon fibre

fabric. The landing gear attachment panel was reinforced with eight plies, to allow the structure

to withstand the landing loads. All other regions, those that experience low in-flight and landing

loads were laid-up with one or two plies, depending on the locations and loading conditions.

During the in-flight analysis step, the carry-through spar attachment and the rear bulkhead pin

attachment locations were found to be the highly stressed. Corresponding Von Mises stress and

in-plane shear stress distributions are illustrated in Figure 7.16.


The Margin of Safety5 (MS) for each of the Von Mises stress and shear stress conditions was

0.450 and 0.104 respectively. In the landing analysis step, the maximum Von Mises stresses

were experienced at the location where the landing gear attachment bracket interfaces with

the fuselage wall, as shown in Figure 7.17. Its MS was determined to be 0.018, the lowest in the

analysis. The shear stresses experienced by the skin upon landing are shown in Figure 7.18. The

corresponding MS was found to be 0.124. Higher MS values (MS > 0.1) at several locations

suggest that the fuselage layup can be optimized further to reduce weight. However, no more

FEA iterations were performed to bring the MS values down at these locations, in order to leave

provisions for the uncertainties associated with the material property formulations and

modelling assumptions. Full scale destructive testing of the fuselage and further design

optimization were recommended for future research.

In both analysis steps, the stresses in the core were considerably lower than the material

strengths, indicating that core failure is unlikely to occur. This is mainly due to the rigid inserts

distributing the applied discrete loads over a larger surface area within the core. Maximum in-

plane strains on the skin, under in-flight and landing loads were also well below the ultimate

strain of the material (0.011). This showed that the linear-elastic FEA assumption is reasonable.

These results are available in Appendix E.

Margin of Safety (MS) = - 1, is a measure of the structural capacity. MS value of zero means Design Load

that the structure will not take any additional loads before it fails. Design optimization work targets to achieve small positive MS values close to zero. Large positive MS values would mean that the structure is overdesigned and negative MS values would mean that the structure will fail before reaching its maximum design load. (Ref: Burr, A and Cheatham, J: Mechanical Design and Analysis, 2nd edition, section 5.2. Prentice-Hall, 1995)


The optimized fuselage structure was estimated to weigh approximately 12.8 lbs. This is 42%

lower compared to its current prototype counterpart, which weighed 22 lbs. The estimates

were based on the surface area and volume of the foam parts used to construct the fuselage

structure. Details on the calculation procedures and the weights estimates of the individual

fuselage panels are also available in Appendix E.

s, Mises SNEG, (fraction = - 1 O). Layer = 1 (Ave;: 7S%) _ - +6.977e+04 B - +3.0OOC+O4

+2.7S0e+04 I Z.SOOc I 04 +2.JS0e+04 + 2 000e+04

I+1.750e+04 +l.sooe+04 + 1.250 e+04 + 1.0OOe+O4 +7.S03e+03 +5.003e+03 +2.503 e+03 +3.615e+00

S, S12 SNEG, (fraction . -1 0), j y e r - 1 (Avg: 75%)

+1.3748+04 +3.0005+0S +2.500S+03 +2.0006+03 +1.5002+03 + i nno*+03 +5.0008+02 +0.0005+00 -5.030e+02 -1.0DOe+03 - l .bjue+iw -2.030e+03 -2.500e+03 -3.030C+03 -1.2S96+04

'• In_Ftight Increment 1: Stap Time => 1.000 Primary var S, Mises

OOft. j Z * . . M M A I W . M . I I I » 1 T". On U 11 U M C O M -

JL i: In Fliqht

Intrenent l:StepTtne = 1X00 Pnmary Var: 3, S12

CMjLeflb Mmi>VEl>*M-i » • » • 6#-I T'.CaZl 11 SI * b><> t*Y«fl'l

Figure 7.16: Von Mises (left) and in-plane shear (right) stresses (psi) in the skin under flight loads


3 M^es C1\£G (fraction a (Avg 75%)

+1 C32e+OS • 3 600C+O4

. +2 7S0C+O4 - +2 cQ3e-KS4 • +2 2S9e+04 - *2 C00e+04 -+1 7S0e+O4

•»1 c 0 9 e + 0 4 +J 250e+O4 • l 0306+04 • 7 501B+03

- •SCOle*03 - +2 SOle+03 - +b £33e-01

Figure 7.17: Von Mises stresses (psi) in the skin during the landing step

S, S12 SNEG, (fracaor = (<\vg 75%)

+ 1 350e+D4 +3 O00e+D3 +2 500e+D3 +2 OO0e+D3 +1 SOOfr+33 + 1 O00e+33 + 5 O00e+D2 +O OOOe+DO -5 DOOe-t-02 - I D00e+03 -1 5006+03 -2.D00C+03 -2 500e+03 -3 DOOe-f-03 -1 2S9e-t-04

1 0), La/er = l

Step Lancing Incremant 1 Step Time — 1 0 0 0 Fnman Var S, S12

OMt jZ<ad> A c * q ^ E u ^ 4 4 Vta»o- t 8 -Z T«. M U 13 SJ 54 E i u - D«f I *IT~—-* IMS

Figure 7.18: Shear stresses (psi) in the skin under landing loads


7.2. Experimental Verification of the FEA results

An experimental study was carried out to verify the selected FEA results. This was required due

to the simplifying assumptions made in the FEA, as explained in section 7.1. The study aimed to

evaluate two major local loading conditions experienced by the fuselage: in-plane bearing and

out-of-plane bending (Figure 7.19). These conditions were present at the location of shear pin

and on the rear bulkhead. Flat sandwich specimens representing these locations on the

fuselage were loaded until failure and the results were compared with the FEA predictions. The

test matrix and the procedure employed in this study along with the results are discussed in the

following sections.

Rigi

Bolted joint

(a) Bearing load (b) Bending load

Figure 7.19: Loading modes chosen for experiments


Sandwich specimen

J insert -mmm-

7.2.1. Test Matrix, Specimen Manufacturing and Test Procedure

The FEA verification test matrix is given in Table 7.6. All specimens were manufactured with

the materials selected for the new fuselage (Chapter 5). The bearing test specimens featured

6.2 lbs/ft3 density PVC foam core having thickness of 0.85 in. and dimensions 6 in. (L) x 6 in.

(W). At the centre of these specimens, the foam core was reinforced with 1 in. diameter FRP

inserts. The specimen thickness was chosen to represent the actual shear pin location in the

new fuselage. Dimensions for the test specimens were chosen based on previous research [58-

60] to avoid edge effects and provide sufficient surface area for clamping of the specimen in the

fixture.

Table 7.6: FEA verification test matrix

Number of specimens

Tested 4 1 3

3 3

Test Type

Bearing Bearing Bearing

Bending Bending

Specimen Dimensions: LxWxT(in.) 6 x 6 x 0.85 6x6x0.85 6 x 6 x 0.85

6x6x0.5 6x6x0.5

Layup

[±45°,0790°, Core]svM [±45°,0790°,0790,Core]SYM [±45°,0790o,0790o,±45°, CoreJsYM

[±45°,0790°, 0790°, Core]SYM [±45°/0790°,0790o

;±45°, Core]SYM

Insert Dimension/Material

l in. diameter GFRP l in. diameter GFRP l in. diameter GFRP

l in. diameter GFRP l in. diameter GFRP


y

->X

Z

• > X

Figure 7.20: Geometry of the sandwich specimen

During specimen manufacture, the foam cores were cut into 6 in. x 6 in. pieces. This was

followed by drilling of 1 in. diameter through-hole at the centre of these foam specimens and

bonding the FRP inserts into the holes using structural epoxy adhesive. These foam specimens

were then laid-up with carbon fibre fabric, vacuum bagged and infused with epoxy resin as per

the conventional VARTM method. All specimens were infused under the same vacuum bag

simultaneously, to avoid any inconsistencies in part quality. After curing, the edges of the

specimens were sanded to create smooth, flat edges to facilitate proper installation in the load

frame. A 0.31 in. diameter through-hole was drilled into each specimen to allow insertion of a

tight-fit aluminum bushing. The specimens were then loaded with 0.25 in. diameter steel pins,


using a hydraulic Material Testing System (MTS) frame (22 kip capacity, Model 647.10A-01, Serial #:

1305166). The bearing test setup in the load frame is shown in Figure 7.21.

(a) Front view (b) Isometric view

Figure 7.21: Bearing test setup in the load frame

Bending test specimens were manufactured with 0.5 in. thick PVC foam core and loaded at 2.75

in. away from the centerline, using 0.375 in. diameter steel pin to represent the engine pin joint

at the fuselage rear bulkhead. These specimens were prepared for testing by following the

same procedures employed for the bearing specimens. The bending test setup in the load frame

is shown in Figure 7.22. All specimens were loaded at a cross-head displacement rate of 0.19 -^—, mm

while the force and displacement data were collected at 10 Hz.


(a) Front view (a) Isometric view

Figure 7.22: Bending test setup in the load frame

7.2.2. FEA Simulations

FE models of the test specimens were created in Abaqus following same procedures employed

in the fuselage FEA. Loads and boundary conditions were applied as shown in Figure 7.23. The

models were meshed uniformly using the optimum element size obtained from the mesh

refinement study. Meshed models are presented in Figure 7.24. First, a range of loads during

which failure might occur, was determined for each layup and specimen type using trial

simulations. This allowed determination of the highest and lowest load to be applied in the

simulations. Then, different FE models were setup for the specimen types discussed in the test

matrix (Table 7.6), each with seven analysis steps. In every analysis step, the applied load was

increased by 50 lbs, starting from the lowest load determined from the trial simulations.


Load

(a) Bearing specimen (b) Bending specimen

Figure 7.23: Abaqus model showing the loads and boundary condition of the specimens

(a) Bearing specimen (b) Bendingspecimen

Figure 7.24: Abaqus FE model of the test specimens


7.2.3. Results

Force-displacement curves obtained from the bearing test specimens are shown in Figure 7.25.

The curves are shifted along the x-axis for clarity. Similar loading profiles of the specimens

suggested consistent specimen behaviour. The failure load for each specimen tested is given in

Table 7.7. Specimen 7 was discarded as it was an obvious outlier in the group of four-ply

specimens. This was due to misaligned installation of the specimen in the test fixture. This

specimen misalignment caused local crumpling of the specimen edges during the test, leading

to premature failure.

Table 7.7: Bearing test failure loads

Specimen ID

SP#1 SP#2

SP#3

SP#4

SP#5

SP#6 gp ff-j

SP#8

Number of Plies

2 2

2

2

4

4

4

3

Load to Failure (lbs)

983.34

926.65

931.62

971.19

1264.16

1196.14

867.719 1069.49


1400

1200

1000

£ 800

o 600

• SP# 1; 2 plies

* SP#5;4plies

• SP#2;2plies

• Sp# 6; 4 plies

SP#3;2plies

^SP#8;3plies

SP#4;2plies

0.05 0.1 0.15 Crosshead Displacement (in)

0.2

Figure 7.25: Bearing test: force-displacement data

Table 7.8 and Figure 7.26 compare average experimental failure loads to those predicted by the

FEA. The experimental failure was defined as the maximum load prior to the abrupt loss of

stiffness in the load-displacement curve. The FEA predicts failure when the in-plane shear stress

on the skin exceeds the material shear strength. The failure modes are compared in Figure 7.27.

For all layups, experimental values were within 10% of the FEA predictions. This correlation

suggests that the assumptions made in the FEA were acceptable for this loading condition.

Table 7.8: Bearing test results: FEA predictions and Experiments

ft of plies

2 3 4

# of specimens tested

4 1 2

Failure load predicted by FEA

(lbs) 983 1154 1202

Average experimental

failure load (lbs) 953 1069 1230

% Difference

3% 8% 2%


I f U U -

1350

1300 -

^ 1 2 5 0 " V)

.0 = i 1200 -•o ra O 1150 a> 3 1100 -'ra "" 1050

1000

950

900

•

•

•

\

i i

• Bearing Test FEA

• Experiment

<> I I

i i

2 3 4

Number of Plies

Figure 7.26: Comparison of the failure loads in bearing test

(a) FEA [25x deformation factor] (b) Experiment [Loaded beyondfailure for clarity]

Figure 7.27: Close-up of the bearing failure mode


FEA of the bending tests showed failure due to shear stress in the skin, as shown in Figure 7.28.

Predicted failure load of the three-ply and four-ply specimens was 300 lbs and 375 lbs,

respectively. The experiment results were rather interesting. Figure 7.29 shows the force-

displacement curves of the specimens tested. The data obtained for specimen 1 (SP# 1) was

non-linear due to the pin bending and plastically deforming as the applied load was increased.

In second test (SP #2) the pin was reinforced to prevent bending, but the resulting curve was

still non-linear though there was no specimen failure. This time it was due to the fixture

bending and deforming at high loads. Thus, in subsequent tests a reinforced fixture and pin

were used. This test process refinement is evidenced through the increasingly linear slopes SP

#1 to SP# 3. The test setup in SP# 3 was most representative of the actual setup on the rear

bulkhead. The rest of the tests were carried out with this setup (Figure 7.22).

In the bending tests, all specimens carried 600 lbs (1650 in.lbs) without failure, which was well

beyond the loads experienced by the fuselage. This showed that the assumptions made in FEA,

specifically the tie constraints between the pin and the sandwich structure, were too

conservative. Perhaps a better correlation would have been obtained between the FEA

predictions and the experiments if the joint configuration was modelled with proper contact

formulation.

The specimens not failing beyond ultimate loads signifies that the limiting factor in the design

was not the insert or the sandwich structure, but the 3/8" bolt itself. Performing extensive FEA

was not the primary scope of this research. Hence, the specimens were saved for later, more


detailed FEA analysis and this research proceeded with manufacturing of the demonstrator

fuselage.

S S I2 SNEG (traction = {AJQ 7 5 M

* 3 0 l o ' O 4 • 1 529*04 *1 260*04 *1 0 l e *04 *7 59o*03

- *5 060*03 *2S3e*03

. *0 0 )*00 - 2 53e>03 - 5 06e*03 -7 59e*03

1019*04 - 1 28a*04 -1 52o*04

• - - 3 20e*O4

ODB Beiw?»ig 4ply oOb Atmqus SlarWatxJ Vwsiot

Slep ln.FlKjW_5 lne*«wn©i« i Step TMPIO - ! 000 Prmnry V I M S S i 2 Defomied Var U Delormabon Seal* Facte*

Figure 7.28: Bending test FEA prediction

• SP# 1; 3 plies • SP#2;3plies - SP#3;3plies

SP#4;4plies < SP#5;4plies *SP#6;4plies

0.2 0.4 0.6 0.8

Crosshead Displacement (in).

1.2

Figure 7.29: Bending test: force-displacement data


CHAPTER 8. FUSELAGE MANUFACTURING

This chapter discusses manufacturing of the demonstrator fuselage using mouldless CCBM

method discussed in Chapter 4. The GeoSurv II fuselage, as shown in Figure 6.11, is a large

component with complex geometry, measuring approximately 44 in. x 15 in. x 13 in. Therefore,

any significant error made during manufacturing of the fuselage would result in an expensive

scrap part. Hence, prior to manufacturing of the full scale fuselage, a sample test section was

manufactured to practice and study the mouldless CCBM techniques. This trial component was

sectioned into small pieces and their cross-sections were examined under the microscope to

assess the quality of the CCBM manufacturing. Then, a series of CCBM experiments were

carried out on flat laminates to determine the optimum spacing between the resin distribution

lines in a CIB infusion. Results from these experiments were utilized to develop a conceptual

manufacturing model in Pro/E, followed by the actual fuselage manufacturing. This work is

discussed in the following sections.

8.1. Sample Section Manufacturing

An H-shaped structure representative of the fuselage rear wall assembly was manufactured

using the CCBM method developed for mouldless manufacturing. This was required to assess

the mouldless CCBM and to identify critical areas that may require process refinement. In this

work, the foam parts shown in Figure 8.1 were machined using a 3-axis Computer Numerical

Control (CNC) router. Rabbet features with male-female profiles were integrated into these

parts to ensure precise-fit assembly of the base plate and side walls. The manufactured foam

part (Figure 8.2 (a)) was laid-up with one ply of carbon fibre fabric and a layer of peel ply.

Subsequently, the sandwich layup was wrapped with a layer of static cling PVC (Type I) film


(Translucent, Low Tack, 0.002" Thick, McMaster Carr Inc.). The CCBM bag was manufactured

over this setup, with inlets, outlets and resin distribution lines positioned as shown in Figure 8.2

(b). The PVC film protects the part from silicone contamination during bag manufacturing while

providing a smooth surface finish to the bag. The chemical resistant nature of the PVC film

allowed the bag to be peeled off easily without using any additional release agents. The low

tack adhesive backing of the PVC film eases the wrapping process, while allowing easy removal

from the setup, after bag manufacturing. The CCBM bag was manufactured over this setup,

following the procedure outlined in Appendix B.

The manufactured CCBM bag (Figure 8.2 (c)) featured two resin distribution lines 3 in. apart,

which distributed the resin from the inlet lines located at the top and bottom surfaces of the

base plate into the part. A layer of 1 in. wide Teflease tape was bonded along the perimeter of

the H profile. Once cured, the bag was cut-opened over the taped surface, using an Olfa® utility

knife. The bag was precisely cut, without damaging the fabric reinforcement, in order to retain

the structural integrity of the component. Upon opening and removing the bag, the PVC film

was peeled from the part. This was followed by re-installing the bag over the part and sealing

the split-ends using disposable sealant tape. Then, the inlet lines were clamped-off and the

outlet lines were connected to vacuum ports. Upon vacuum application, it was noticed that the

side walls warped under the applied pressure, due to the unsupported nature of the part

geometry. This problem was solved with the insertion of rigid spacers between the side walls.

Once the setup was complete, resin was introduced into the part via the inlet line. The infusion

of this test component took approximately 6 minutes. Following the infusion, the inlet lines


were attached to vacuum ports to remove any excess resin from the part. The part was cured

under vacuum pressure for 24 hours.

Figure 8.3 shows the cured component before and after removal of the peel ply. There was

some residue of cured resin left in the resin distribution channels and inlet lines, which is

normal in CIB infusion. As the peel ply was removed, the cured resin channels snapped off from

the part, leaving negligible marks on the surface. Significant improvement in the surface quality

was observed around the corner regions, compared to the current fuselage [3]. The part

retained nicely formed corner radii with very few resin starved regions. This improvement in

quality can be attributed to the form-fitted nature of the CCBM bags.

Figure 8.1: Geometry of the test article

In order to assess the quality of the sandwich structure manufactured by the mouldless CCBM

method, the test article was sectioned into small pieces and their cross-sections were examined

under the microscope. In this analysis, images taken from the test specimens were compared


against the images of coupons manufactured using conventional VARTM method. A sample

comparison is shown Figure 8.4. The partially open cells at the skin-core interface were found

to be filled with resin and the trends were comparable between the VARTM and the CCBM

specimens. This is required for good adhesion and proper load transfer between the skin and

the core. More microscopic images of the VARTM and the CCBM specimens are available in

Appendix F.

(a) Sample H-Section (b) CCBM Bag Manufacturing (c) CCBM-CIB Infusion

Figure 8.2: Important features of mouldless CCBM setup

(a) Manufactured test article (b) Manufactured priorto removal of peel ply test article

Figure 8.3: Manufactured component


Figure 8.4: Bondline comparison of VARTM and CCBM manufactured sandwich coupons

8.2.CCBM Experiments and Fuselage Manufacturing Model

In this work, an appropriate spacing between resin distribution lines was determined based on

flat panel experiments. Results from these experiments were combined with the knowledge

acquired from the current prototype manufacturing (Maley, [3]), to develop a conceptual

fuselage manufacturing model. The objective of this work was not shortening the infusion time,

but rather creating a robust and efficient manufacturing setup.

Two flat panel CIB-CCBM experiments were carried out to determine the optimum separation

between the resin distribution lines. Each CCBM bag had a resin inlet, a vacuum outlet and two

resin distribution lines. Important parameters of the test setup are shown in Figure 8.5. Setups

of the two experiments were identical except for the separation between the resin distribution

lines (variable (d) in Figure 8.5 ), which were 6 in. for the first experiment and 10 in. for the

second. The distance between the resin inlet and vacuum outlet was set approximately at 24 in.

This distance was chosen to represent the maximum distance from the centreline of the

fuselage to the centre of the side wall. The resin distribution lines moulded into the bag were

kept 3 in. short of the vacuum outlet line. If this separation was not present, the resin would


travel preferentially from the inlet to outlet rather than infusing the part. The infusions were

carried out on a flat tool surface with four ply carbon fibre fabric preform, at [(±45°)2, (0790°)2]

layup.

The CCBM setup that featured 6 in. separation between the resin distribution lines infused the

part in approximately 13 minutes, while the other setup took 25 minutes for complete infusion.

From manufacturing of the test section, it was found that 3 in. spacing between the resin

distribution lines would require 6 minutes for infusion. From these results, it was decided that

the maximum separation between any two resin distribution lines on the fuselage should be no

more than 8 in. This would result in infusion time of approximately 20 minutes. Though the

infusion time associated with this option was not the shortest, the infusion can be

accomplished with fewer resin distribution lines. This reduces the complexity of the

manufacturing setup and, in turn, makes the additional infusion time tolerable.

Figure 8.5: CCBM experiment setup


Following the experiments, a conceptual CCBM manufacturing model was developed using

Pro/E. The final manufacturing model shown in Figure 8.6 had resin distribution lines at various

spacings along the fuselage walls. Areas of single ply layups had resin distribution lines 8 in.

apart. As the ply count increased towards the rear bulkhead and the front bulkhead of the

fuselage, this spacing was reduced to facilitate simultaneous infusion, as shown in Figure 8.6.

Resin inlet lines were positioned along the centerline of the base plate and over the landing

gear attachment panel, while vacuum outlet lines were placed around the perimeter of the

fuselage. Approximately 2.5 in. distance was kept between the outlet line and the resin

distribution lines.


Isometric View Bottom Section

Figure 8.6: Conceptual CCBM Manufacturing Model


8.3. Fuselage Manufacturing Fuselage manufacturing involved three primary steps: foam core preparation and layup, bag

manufacturing and CCBM infusion. The manufacturing sequence is illustrated in Figure 8.7.

Step 1: Core Preparation/ Layup Step 2: Bag Manufacturing

Manufactured Fuselage Step 3: CCBM Setup/Infusion

Figure 8.7: Mouldless CCBM Process

Manufacturing of the fuselage began with preparation of the foam core. In this process,

required foam parts were machined using a 3 axis-CNC router. The machining process is shown

in Figure 8.8 and the machined parts are shown in Figure 8.9. Then rigid FRP inserts (Figure

8.10), were cut to size from FRP rods, sandblasted and bonded into the foam parts using

PTM&W Inc. ES 6220 epoxy (Pot life 4-6 minutes, cure time 15 minutes). The foam parts were

then bonded together to build the fuselage structure. A custom designed and built wooden

assembly jig, shown in Figure 8.11, was used to hold the foam parts together during the

bonding process. The assembly jig and the alignment features included in the foam parts

(Figure 8.12) served to develop a robust core structure of the fuselage in a cost efficient and


repeatable manner. Following assembly of the foam parts, the fuselage was laid-up with carbon

fibre fabric (BGF Style# 94132), in accordance to the optimized layup determined from the FEA

(Figure 7.15). A layer of thin fibreglass fabric (BGF Style# 106) was laid-up around the outer

surface of the fuselage to provide fine surface finish and thereby minimize the surface

preparation required for painting. The fuselage assembly was wrapped tightly with a layer of

peel ply to create uniform surface finish. During layup, extra care was taken to ensure accurate

fabric orientations and minimum overlaps on the OML. Additionally, it was ensured that all

fabric and peel ply layers were tacked onto the foam core and conformed around the corner

regions. This effort was important to generate good surface finish and minimize resin starved

regions in the fuselage.

Foam Part

Figure 8.8: Machining of the foam parts on the CNC router table


LAvionicsMount Base Plate

<o> ttl

. . • , ! > -

Side Walls * m

Front Walls Front Bulkhead

1 Fairings

:igyr<e 8.9: Feam parts reqyiredl for tfyselag® manyfacSyrifiig

W inserts for ear trough spas' 2*0 g'l V> I


Alignment Fences

Bonding Fixture

Demonstration of the bonding strategy

Figyre 8.11: (Bonding of foam parts in the tenure

v Male-female extrusions

k_ \ ' • L L ^

J. Alignment Shelves

"^'J^| Inserts used as alignment dowels

r^flB Figyre 8.12: Featyres inctoded in tthe foam parts to facilitate precise assembly

CHAPTER 8. FUSELAGE MANUFACTURING

In the next stage of manufacturing, the fuselage core assembly was used as the mould to

manufacture the CCBM bag. The techniques employed in bag fabrication were identical to the

CCBM trial section manufacturing described in section 8.1. The manufactured CCBM bag was

equipped with resin inlets, vacuum outlets and resin distribution channels as illustrated in

Figure 8.6. Upon curing, the bag was cut open along the taped surface and liquid mould release

was sprayed over the inside surface of the bag. This optional step was included in the

manufacturing process to add extra protection to the bag and enhance its self-release

capability.

In the final step, the manufactured CCBM bag was installed back onto the fuselage layup and

sealed with disposable sealant tape. The outlet lines were attached to the vacuum ports. Upon

application of vacuum pressure, the fuselage walls were found to warp inward. The walls were

supported using spacers made from turn buckles with swivel feet (Figure 8.13), to correct this

warpage. Once the setup was complete, a vacuum integrity check was performed, where 1

inHg/min. was considered an acceptable vacuum loss. The CCBM setup had a vacuum loss of

0.3 inHg/min., a significant improvement upon the previous mouldless VARTM setup (Maley,

[3]).


Rigid Spacers

Figure 8.13: Mouldless CCBM setup

After verifying the vacuum integrity of the setup, the fuselage was infused with PT&W 2712

epoxy. Though the infusion time predicted based on the experiments (section 8.2) was 20

minutes, the actual infusion was completed in approximately 45 minutes. This was primarily

due to the resin shortage, which occurred about 10 minutes into the infusion. When this was

noticed, the resin feed lines were clamped off, a new batch of resin was mixed and the infusion

was resumed after 10 minutes. This slowed down the infusion process and resulted in small

resin starved regions in the bottom base plate section, as shown in Figure 8.14. The flow

behaviour during the infusion (Figure 8.7-Step 3) was similar to that observed in the flat panel

CCBM experiments, where the resin travelled preferentially through the distribution lines and

then spread into the fabric.


Figure 8.14: Resin starved regions observed during the infusion

Following the infusion, inlet lines were attached to the vacuum ports to remove any excess

resin from the part. The part was left to cure under vacuum pressure, for 24 hours. The

manufactured fuselage is shown in Figure 8.7.


CHAPTER 9. MANUFACTURING RESULTS

The manufactured fuselage was visually inspected to assess the surface quality. The dimensions

of the new fuselage were measured and compared against the current fuselage manufactured

by mouldless VARTM (Maley, [3]). The weight of the new fuselage was measured and assessed

against the FEA predictions. Finally the PVA was revisited to draw conclusions on the process

viability. This work is described in the following sections.

9.1. Surface Finish, Weight and Tolerances

The manufactured fuselage had a fine peel ply finish with carbon-epoxy skin tightly conformed

around the corners. Some of these surfaces are highlighted in Figure 9.1. With the exception of

the small resin starved regions found in the base plate, the fuselage was fully infused and

showed smooth surface texture. Overall surface finish was better than that observed in

previous manufacturing attempts.

Figure 9.1: Fuselage Manufactured by mouldless CCBM


There were some issues with the tolerances, although the modified CCBM setup alleviated the

wall distortion observed in the previous mouldless VARTM. The side walls near the top-rear

section of the fuselage were found to be deflected inward, as shown in Figure 9.1. This

deflection was a direct consequence of one turnbuckle spacer falling off, leaving part of the

fuselage walls unsupported during the cure. All other walls remained supported throughout the

entire process. To better characterize the dimensional tolerances, the OML of the

manufactured fuselage was profiled. In this work, the outer dimensions of the fuselage were

measured using rulers, to an accuracy of ±1/16 in. Measurements were taken at a sufficient

number of points to establish an accurate profile of the final shape. Detailed profiling can be

found in Appendix G [61].

Deviations from the target dimensions were below 0.09 in. in most regions, with the exception

of the rear walls on the top half of the 'H' structure and front walls on the bottom half of the 'H'

structure. At these locations the maximum deviations were found to be 0.4 in. and 0.23 in.

respectively. Compared to the previously implemented mouldless VARTM, which resulted in

surface variations of up to 0.25 in. along most fuselage walls, the tolerances achieved with this

manufacturing method better complied with the design targets. This is shown in Table 9.1.

However, there is still a need for improved techniques for controlling part dimensions in

mouldless manufacturing. Better dimension control can perhaps be achieved with the use of a

connected fixture rather than individual spacers (similar to the VARTM setup, [3]). If individual

spacers are used, they need to be embedded into the vacuum bag to ensure more accurate

positioning.


Table 9.1: Comparison of major dimensions: fuselage design vs. current and new fuselages

Measurement Location

Rear width

Front width

Length (Centerline) -Top

Length (Left) - Top

Length (Right) - Top

Length (Centerline) - Bottom

Length (Left) - Bottom

Length (Right) - Bottom

Length - Fairing to Fairing

Design (in.)

13.34

10.93

43.03

43.08

43.08

44.15

43.02

43.02

15.55

IVIeasured dimensions

(new fuselage) (in.± 1/16 in.)

13.31

10.88

42.97

43.06

43.03

44.19

43.06

43.06

15.47

Deviation in the new fuselage

(in.)

0.03

0.05

0.06

0.02

0.05

-0.04

-0.04 -0.04

0.09

Deviation in the current fuselage

(in.)

-0.06

-0.10

0.20

0.01

0.05

0.31

0.07

0.07

0.05

The manufactured fuselage weighs approximately 14.6 lbs. In order to compensate for the

material to be removed during the installation of the carry-through spar, 0.5 lbs was deducted

from the measured value. Possible weight reduction offered by drilling holes at the locations of

inserts was ignored. Thus the effective weight of the new fuselage was estimated to be 14.1 lbs.

Compared to the current fuselage, the new fuselage offered a total weight saving of 7.9 lbs,

which was approximately 36% of the fuselage weight and 4% of the entire aircraft weight. The

actual weight was close to the theoretical weight estimated using the FEA, which is summarized

in Table 9.2. The Fibre volume fraction (Vf) of the manufactured fuselage was estimated to be

51%. The value of vywas not determined using an ASTM standard, but rather an approximate

value was estimated based on theory, measured weights and assuming zero void content.

Detailed Vf calculations are provided in Appendix H.


Table 9.2: Comparison of the actual weight and predicted weight of the new fuselage

FUSELAGE WEIGHT SUMMARY Current fuselage weight Predicted weight of the new fuselage based on the FEA Weight of the new fuselage as manufactured

Predicted weight saving

Actual weight saving

22 lbs 12.8 lbs

14.1 lbs 9.2 lbs

(42% of the fuselage, 5% of the aircraft)

7.9 lbs (36% of the fuselage,

4% of the aircraft)

9.2. Process Viability

During manufacturing of the fuselage, it was realized that approximately one week was

required to complete the CCBM bag6. This contradicted the initial PVA estimates, which

accounted only two days for bag manufacturing. Though the CCBM bag could be manufactured

within two days using fast-cure, sprayable CCBM systems, it was in the best interest of this

research to determine whether the CCBM method employed was cost-effective for producing

components in quantities below 10 units. Hence, the PVA was revised to account for a week of

bag manufacturing, and the resulting costs at labour rate of 40$/hr are shown in Figure 9.2.

From these results, it is apparent that the method employed in this research, CCBM III, would

be an attractive choice for manufacturing more than six fuselages. Thus, the process utilized is

viable for mouldless manufacturing of components in low production quantities.

6 Note that in this bag manufacture, some effort was related to the development.


$5,000

$4,500

$4,000

$3,500

» $3,000 • -to

o $2,500 • -

| $2,000 •

$1,500 -

$1,000 -

$500 • $0

- • - VARTM

-•-CCBM-I

- - C C B M II

- * -CCBM III

4 6

Number of Parts

10

Figure 9.2: Revised PVA cost estimates based on actual labour required for fuselage manufacturing (labour rate $40/hr)


CHAPTER 10. CONCLUSIONS

The conclusions drawn from this research and recommendations for future work are discussed

in this chapter.

10.1. Conclusions

> Based on an in-depth review of the current LCM processes, CCBM was selected for

mouldless manufacturing of foam-core composite sandwich components. The initial

material and labour cost associated with CCBM is slightly higher than that of

conventional VARTM, but this added cost comes with the benefits of improved process

robustness, repeatability and part quality. A series of flat-panel experiments followed by

a PVA carried out in this research demonstrated that mouldless CCBM with CIB infusion

is a viable option for low volume (<10) production of large, complex components.

> PVA is an effective tool for determining the feasibility of a process. When several

process variants exist, PVA can be used to select the most viable option.

> Effective application of DFM and near-net-shape manufacturing principles improves the

part quality, tolerances and reduces the overall cost of production, in mouldless

manufacturing.

> Simplified FEA techniques are fast and effective in determining the optimum layup of a

composite structure.

> Weight savings of 8 lbs (36% as compared to the current fuselage prototype) were

achieved on the GeoSurv II fuselage using design optimization based on DFM principles

and FEA, supported by coupon tests.


> To facilitate the transfer of discrete loads into a sandwich structure, rigid inserts

embedded into the foam core prior to resin infusion are a robust solution.

Subsequently, holes can be drilled through these local "hard points" and bolted joints

can be created. Such joints are typically lighter and more effective in distributing the

loads into the surrounding structure as compared to those created by bonding rigid

inserts into the manufactured component.

> A new full-scale GeoSurv II fuselage was manufactured to near-net-shape in a single step

infusion using an improved mouldless CCBM method. Although the new fuselage

showed improved surface quality and dimensional tolerances compared to the current

fuselage, some filling and sanding is still required to bring the tolerances closer to the

design specifications. However, achieving near-net-shape tolerances with mouldless

CCBM is not far from reach.

> This research has contributed to several important aspects of low cost composite

structures, including structural design with material selection, structural optimization,

design for manufacture (DFM), process selection, process value analysis (PVA),

manufacturing process planning and development.

10.2. Recommendations for Future Work

The dimensional tolerances of mouldless CCBM manufacturing can be improved by embedding

rigid spacers into the CCBM bag or with the use of accurate fixtures. In the future, more

fuselages will be manufactured using the CCBM bag employed in this research, but with a

modified technique for part shape retention. Results from future manufacturing shall be used

to verify the process repeatability for complex-shape components.


Future manufacturing should also consider using fast-cure, sprayable CCBM systems to improve

the efficiency of bag manufacturing. If not constrained by the design limitations, the foam core

should be designed to withstand the applied vacuum. Thus, the need for additional support

fixtures for part shape retention can be eliminated.

Accurate modelling of the bolted joints using a proper contact algorithm is recommended to

obtain better correlation between FEA results and experiments, particularly for the out-of-

plane bending loads. Additionally, properties of the new matrix and reinforcement materials

should be characterized using coupon testing and the material properties used in the FEA of

this research should be revised to improve the results. Future FEA work should also consider

simulating and optimizing the fuselage under fatigue load and impact loads.

Assumptions used in the FEA were conservative and the MS values were somewhat high at

several locations, which mean that there is still room for optimizing the fuselage layup and

further reducing weight. Future work should consider testing the new fuselage to failure and

optimizing the layup based on the test results and improved FEA methods.

Redesign work in the future should also consider simplifying the geometry to ease the fabric

layup procedures. Additionally, the FRP inserts used in this research have directional properties,

which are undesirable in the presence of multi-axial loads. Thus, future redesign work should

look into isotropic inserts. To further optimize these joints, different insert and joint

configurations should be developed and compared against the current joint design.


Detailed research on flow modelling and developing strategies for optimizing the infusion setup

would further improve the capabilities to manufacture arbitrary (complex) structures using

CCBM.


REFERENCES

1. "Sandwich Panels", Lecture Notes for AERO 4608, Composite Materials, Lecture 14e-Design, Department of Mechanical and Aerospace Engineering, Carleton University, 2008.

2. Diab Inc., Foam Core Materials in the Marine Industry, [Online- Technical Bulletins], 2009, [Cited May 2010] Available: http://www.diaberoup.eom/americas/u literature/u pdf files/u bul pdf/Foam Core Marine TB.pdf.

3. Maley, A.J, "An investigation into low-cost manufacturing of carbon epoxy composites and a novel mouldless technique using the Vacuum Assisted Resin Transfer Moulding (VARTM) method", MASc Thesis, Department of Mechanical and Aerospace Engineering, Carleton University, Ottawa, ON, 2008.

4. Mahendran, M. et al., "Feasibility Study of Closed Cavity Bag Moulding for Novel Mouldless Manufacturing of Carbon-Epoxy Composites", presented at the Canadian Aeronautics and Space Institute AERO '09 Conference- 20th Aerospace Structures and Materials Symposium, Ottawa, ON, 2009.

5. Carleton University Department of Mechanical and Aerospace Engineering and Sander Geophysics Limited. GeoSurv II Unmanned Aerial Vehicle: System Requirements Document (UAV-SRD), Rev. E, 31 March, 2008.

6. Summerscales, J, Searle, T.J, "Low-pressure (vacuum infusion) techniques for moulding large composite structures", Proc. IMechE Vol. 219 Part L: J. Materials: Design and Applications, 07 September 2004.

7. Beckwith, S.W., and Hyland, C.R, "Resin Transfer Moulding: A decade of Technology Advances", SAMPE Journal, Vol. 34, No. 6, November/December, 1998.

8. Rudd, CD. et al., "Liquid Moulding Technologies, Resin Transfer Moulding, Structural Reaction Injection Moulding and Related Processing Techniques", Woodhead Publishing, 1997, pp. 1-35.

9. J.H.A. van der Woude and E.L. Lawton, "Composite Design and Engineering, Fibreglass and Glass Technology", ISBN 978-1-4419-0735-6. Springer Science+Business Media, LLC, 2010, pp. 125-173.

10. Summerscales, J., "Composites Design and Manufacture- Composites Manufacturing Processes" University of Plymouth- Advanced Composites Manufacturing Centre, MATS 324 Lecture Notes, [Online Version Cited May 2010], Available: http://www.tech.plym.ac.uk/sme/MATS324/MATS324C.htm.

11. Potter, K., "Resin Transfer Moulding", Springer-Verlag, 1997.

12. Pant, S. et al., "Characterization of Double-Bagging Effects on 1-D Permeability for Vacuum Assisted Resin Transfer Moulding (VARTM) Process", presented at the Society for the Advancement of Materials and Process Engineering- SAMPE 2010 Conference, Seattle, WA, May 2010.

13. Mahendran, M. et al., "Feasibility Study of Closed Cavity Bag Moulding (CCBM) for Novel Mouldless Manufacturing of Carbon-Epoxy Composites" presented at Canadian Aeronautics and Space Institute AERO'09 Conference, 20th Aerospace Structures and Materials Symposium, Ottawa, ON, 2009.

14. Seemann, WH. "Unitary Vacuum Bag for Forming Fibre Reinforced Composite Articles". US Patent No. 5, 316, 462; May 21,1994.

15. DAI J. et al., "A Comparative Study of Vacuum-Assisted Resin Transfer Molding (VARTM) for Sandwich Panels", Polymer Composites, Vol. 24, No. 6, December 2003.

REFERENCES 137

http://www.diaberoup.eom/americas/u

http://www.tech.plym.ac.uk/sme/MATS324/MATS324C.htm

16 Seemann, WH "Plastic transfer molding techniques for the production of fibre reinforced plastic structures" US Patent No 4,902, 215,1990

17 LI, W etal, "Process and Performance Evaluation of the Vacuum-Assisted Process" Journal of Composite Materials, Vol 38, No 20, 2004

18 Allende, Metal, "Experimental and Numerical Analysis of Flow Behavior in the FASTRAC Liquid Composite Manufacturing Process", Polymer Composites, Vol 25, No 4, August, 2004

19 Niggemann, C et al, "Experimental Investigation of the Controlled Atmospheric Pressure Resin Infusion (CAPRI) Process"', Journal of Composite Materials, 42, SAGE Publications, 2008

20 Takeda, fetal, "Research in the Application of the VaRTM Technique to the Fabrication of Primary Aircraft Composite Structures", Mitsubishi Heavy Industries, Ltd Technical Review Vol 42 No 5, December, 2005

21 Klemeberg, M etal, "Cost Effective CFRP-Fuselage Manufacturing with Liquid Resin Infusion (LRI) -Technologies", Workshop at German Aerospace Centre (DLR) on Final Project of Black Fuselage, Braunschweig, Germany, 2002

22 Louderback et al, "High Performance Infusion System for VARTM Fabrication", U S Patent 6 964 561 B2 November 15, 2005

23 Gibson, R F and Ayormde, E O, "Vibration-Assisted Liquid Composite Moulding", ANTEC Conference Proceedings, Society of Plastics Engineers, Brookfield, CT, vol 2, pp 1544-1547, 2004

24 Arctek Inc, "Closed Cavity Bag Moulding Multi-Port Infusion", CCBM Training Manual, Rev 26 November 1999

25 Fink B K, et al, "Co-Injection Resin Transfer Moulding of Vinyl-Ester and Phenolic Composites", The U S Army Research Laboratory Technical Report, ARL-TR-2150, Accession Number ADA 373528, January 2000

26 Bottler, O et a l , EC-HLM Honeycomb Liquid Moulding, [Online Document] EURO Composites, 2008, [Cited May 2010] Available http //www euro composites com/SiteCollectionDocuments/EC HLM EN pdf

27 Magnum Venus Plastech Ltd , Light RTM (LRTM), [Online], 2007, [Cited September 2010], Available http //www plastech co uk/Mtlrtm html

28 JHM Technologies Inc, RTM Light- The Lower Cost Alternative Takes the Lead in the Closed Moulding Industry, [Online], 2009, [Cited September 2010], Available http //www rtmcomposites com/lrtm html

29 Thagard, J R , et al, "Resin Infusion Between Double Flexible Tooling Evaluation of Process Parameters", Journal of Reinforced Plastics and Composites 23,1767, SAGE Publications, 2004

30 Ruiz, E and Trochu, F, "Flexible Injection A Novel LCM Technology for Low Cost Manufacturing of High Performance Composites Part I Experimental Investigation", 9th International Conference on Flow Processes in Composite Materials (FPCM 9), Montreal, Canada, July 8-10, 2008

31 Oj, B etal, "A Resin Film Infusion Process for Manufacture of Advanced Composite Structures", Composite Structures- 47, pp 471-476,1999

REFERENCES 138

32 Airtech Inc., Vacuum Baggin Techniques, [Online] © 2008, 2009, 2010 - AIRTECH Europe Sari, [Cited September 2010] Available: http //catalogue airtech lu/product php?product id=355

33. Airtech Advanced Materials Group, Airtech Europe S. A., Rubber Silicone Seals, [Online Catalogue], 2008, [Cited 12Jan09], Available: http://cataloeue.airtech.lu/product.php7product id=29&lane=EN.

34. Aaron Miller, CCBM, E-mail communications with Aaron Miller from Composites Canada, Mario Mahendran, Ottawa, 2009.

35. Arctek Inc. CCBM Start-up Kit, [Cited October, 2010], Available: http / / w w w arctekmfusion com/Howtogetstarted htm

36. Crow, K., Value Analysis and Functional Analysis System Technique, DRM Associates, [Online Document], [Cited March 12, 2009], Available: http / / w w w npd-solutions com/va html

37. Lisa Somanchi, et. al., Value Analysis: Overview, The Quality Portal [Online Document], September 09, 2008, [Cited March 12, 2009], Available: http //thequahtyportal com/articles/value htm

38. Diab Inc., Sandwich Concept, Diab Manuals- Diab Sandwich HandbookDiab, [Online Document], [Cited Apr. 2010], Available: h t t p / / w w w diabgroup com/europe/hterature/e pdf files/man pdf/sandwich hb pdf

39. Allen, H.G., "Analysis and Design of Structural Sandwich Panels", Pergamon Press, New York, 1969.

40. Plantema, F. J., "Sandwich construction; the bending and buckling sandwich beams, plates, and shells", Wiley, New York, 1966.

41 . Daniel I.M., et. al., "Major Accomplishments in Composite Materials and Sandwich Structures", An Anthology of ONR Sponsored Research, Springer, New York, 2009.

42. Black S., Getting to the core of Composite Laminates, Composites Technology, Gardner Publications, Inc, [Online Document], 01Oct2003, [Cited 25Jul2008], Available:http / / w w w compositesworld com/articles/getting-to-the-core-of-composite-laminates aspx

43. Nida-Core Corporation, BalsaLite Quality Coated Balsa Core,[Online Document], 2008, [Cited July 2008], Available: http / /www nida-core com/pdfs/pds/nidacore/pds balsalite pdf

44. SP Systems, Core Materials in Polymeric Composites, AZo Journal of Materials Online, AZoM™, [Online Document], 2009, [Cited Aug08], Available: http / / w w w azom com/details asp?ArticlelD=1092

45 Alcan Composites, Foam - AIREX° PXc - Fibre-Reinforced Structural Foam, I.C.A.R.O.H. GmbH/Alcan Composites, [Online Document], 2008, [Cited 27Jul2008], Available: h t t p / / w w w dibond com/alcan/acsites nsf/pages accm3 en/index htmiQpen&p=prod foam pxc&m=4&

type= htm

46. Alcan Composites, Foam - AIREX' PXw - Fibre-Reinforced Structural Foam, I.C.A.R.O.H. GmbH/Alcan Composites, [Online Document], 2008, [Cited 27Jul2008], Available: http / /www dibond com/alcan/acsites nsf/pages accm3 en/index htm |Open&p=prod foam pxw&m=4&

type= htm

47. Nida-Core Corporation, Structural Honeycombs-Foam Filled, [Online Document], 2008, [Cited Aug08], Available: http / / w w w nida-core com/english/mdaprod honey foam htm

139

http://cataloeue.airtech.lu/product.php7product

48. MGI Inc., MiKor Foam Filled Honeycomb, [Online Document], [Cited July 2008], Available: http://www.mgicanada.com/honevcomb.htm.

49. Cochrane D., "Fuselage Finite Element Analysis (FEA)", DR-87-05, Carleton University GeoSurv II Unmanned Aircraft System (UAS) project, 24 March, 2008.

50. Corbett, J.,et. al.. Design for Manufacture: Strategies, Principles and Techniques, Addison-Wesley Publishing Company, Ontario, 1991.

51. Deshpande V.S. and Feleck N.A., "Multi-Axial Yield Behaviour of Polymer Foams", Acta Mater. 49, pp. 1859-1866, 2001.

52. Russell Elkin, PVC Foam Properties, E-mail communications with Russell Elkin, Senior Technical Service Engineer at 3A Composites USA, Mario Mahendran, Ottawa, 2010.

53. Hart-Smith, L. J., 'The ten-percent rule for preliminary sizing of fibrous composite structures", Weight Engineering, vol. 52, no. 2, p. 29-45,1992.

54. Naik R.A., "Failure Analysis of Woven and Braided Fibre Reinforced Composites", NASA Contractor Report 194981, Contract NASI-19708, Notational Aeronautics and Space Administration, Langley Research Centre, Hampton, Virginia, 23681-0001, September 1994.

55. K. Suraweera. V-n Diagram for 150 lb Aircraft and Lift Distribution on a Vertical Tail. Design Report 63-10A. Course AERO 4907. 05 April, 2007.

56. Zakurdaev, A., Update Load Analysis and Methodology, DR 83-02, Carleton University GeoSurv II Unmanned Aircraft System (UAS) project, 31 March, 2008.

57. Buschinelli, M., "Fuselage Loads Summary", DR-117-17, Carleton University GeoSurv II Unmanned Aircraft System (UAS) project, 08 December, 2009.

58. Karakuzu, R. et al., "Failure analysis of woven laminated glass-vinylester composites with pin-loaded hole", Composite Structures, 72; pp. 27-32, 2006.

59. Murat, B. and Sayman, O., Failure analysis of pin-loaded aluminum-glass-epoxy sandwich composite plates, Composites Science and Technology; 63 pp. 727-737, 2003.

60. Baba Okutan, B., Behavior of Pin-loaded Laminated Composites, Experimental Mechanics46: pp. 589-600, 2006.

61. Teutsh, J., "DR137-09: Profiling the Final GeoSurv II Fuselage" Carleton University Aerospace Engineering, Ottawa 26 October 2010.

REFERENCES 140

http://www.mgicanada.com/honevcomb.htm

APPENDICES

Appendix A: Manufacturing Supplies This appendix contains details on the products used in the manufacturing trials and experiments of this research.

Matrix. Reinforcement and Core Materials:

> Infusion Epoxy: PTM&W Inc. PR 2712 [1],[2] > Infusion Epoxy: API SC 780 [3] > Epoxy Adhesive (Fast Cure): PTM&W Inc. ES 6220 [1],[2] > Carbon Fibre Fabric: BGF Style# 94132-3k, 4H satin [1],[4] > Carbon Fibre Fabric: Hexcel AGP 370- 6k, 5H satin [5] > E glass Fabric: BGF Style#106, Ik plain weave [1],[4] > Foam Core: Airex C- structural PVC foam [6] > Inserts: Fibreglass rods [7]

VARTM Supplies:

> Mould Sealant Epoxy - West System Epoxy: Part A105 and Part B 205 or 206 [1],[8] > Paste Mould Release: TR 104 High Temperature [1], [9] > Spray Mould Release: MS-122 AD: PTFE release agent, dry lubricant [10] > Distribution Medium - Resinflow 60 [1],[11] > Vacuum Bag - Strechlon 200, Stretchlon 800 [1], [11] > Sealant Tape - AT-200Y [1], [11] > Breather - Econoweave 44 [1], [11] > Peel Ply - Econoply J [1], [11] > Tubes, Fittings and other consumables- Composites Canada [1], [11]

CCBM Supplies:

> Arctek CCBM system: Progress Plastics and Compounds, Mississauga, Ontario [12] > Teflease tape [1] > 1/8 in. diameter wire wax [13] > Static cling PVC (Type I) film: Translucent, low tack, .002" thick, 12" width, 12'L [7] > Double-sided cloth tape: (Item #. 76125A21- Mc-Master Carr) [7]

Suppliers/Distributors:

[I] Composites Canada, http://www.compositescanada.com/index.php [2] PTM&W Industries Inc., http://www.ptm-w.com/ [3] Applied Poleramic Inc., www.appliedpoleramic.com [4] BGF Industries Inc.. http://www.bgf.com/ [5] Hexcel Corporation, www.Hexcel.com [6] 3a Composites, http://www.corematerials.3acomposites.com/home html?L=l [7] Mc-Master Carr, http //www.mcmaster,com/# [8] West System, www.westsvstem.com [9] TR Industries, http://www.tnndustries.com/ [10] Miller-Stephenson Company Inc, http://www.miller-stephenson.com/ [ I I ] Airtech International, www.airtechonline.com [12] Arctek Inc., http.//www.arctekinfusion.com/What%20is%20CCBM.htm [13] Kindt-Collins Company LLC, http://www.kindt-coilins.com/

Appendix A: Manufacturing Supplies 141

http://www.compositescanada.com/index.php

http://www.ptm-w.com/

http://www.appliedpoleramic.com

http://www.bgf.com/

http://www.Hexcel.com

http://www.corematerials.3acomposites.com/home

http://www.mcmaster,com/%23

http://www.westsvstem.com

http://www.tnndustries.com/

http://www.miller-stephenson.com/

http://www.airtechonline.com

http://http.//www.arctekinfusion.com/What%20is%20CCBM.htm

http://www.kindt-coilins.com/

Appendix B: CCBM Bag Manufacturing Procedure

CCBM Manufacturing Trial #1 & 2

1. Prepare the glass tool by cleaning it with Spray-Nine multipurpose cleaner and accurately draw lines to indicate where the seal and resin inlet/outlet lines are to be located ( Figure Bl ) .

Note: The quality of the bag depends on the surface quality of the tool. Fill-in any rough spots or dents on the tool with modelling clay until a smooth surface is achieved.

wn to indicate the t of the seal and resin t lines

Figure B 1: Tool Preparation

2. Apply three coats of TR-104 High Temperature Mould Release Wax over the surface on which the bag is to be fabricated.

Note: Any non-silicone based mould release in liquid or paste form can be used in this step.

3. Slice a 0.5 in. diameter poly tubing in two halves and secure it on the tool using double-sided tape, at the appropriate location of the inlet/outlet lines (Figure B 2).

Flash breaker 1 Tape

Resin inlet/Outlet Lines

Proflex NS Silicone bead

Waxed Glass Tool

Figure B 2: Applying silicone over the prepared tool


4. Apply release wax over the poly tubing to ensure easy removal after the bag is cured. 5. Place Airtech Flashbreaker tape upside down along the lines where the bag is to be

sealed, using Airtec 2 fast cure spray adhesive. Note: The use of Airtec 2 spray adhesive to hold the Flashbreaker tape in place contaminated the Flashbreaker tape, causing the sealant tape to permanently adhere to its surface. Hence, in trial # 2, double-sided tape was used to hold the Flashbreaker tape onto the mould.

6. Install 850 g Proflex NS Silicone cartridge in a caulking gun and apply approximately 0.25 in. diameter bead of silicone over the tooling surface, in a pattern as shown in Figure B 2.

7. Using econo-bristle brush or plastic squeegees smear the silicone across the tool along the direction specified in Figure B 2, to create the first layer of silicone.

Note: Care must be taken to brush the silicone over the tool as uniformly as possible. Avoid building up this layer over 0.02 in. thick, as it will take longer to cure.

8. Let the first silicone layer completely cure (approximately 1 hour) in air/moisture and apply another layer of silicone over it. This layer of silicone is to balance out any thickness variations from the first layer and also to build up the thickness of the bag.

9. Once the second layer of silicone is cured, apply a very thin layer of silicone over the bag surface and while the silicone layer is wet tack the Confortex fabric onto the bag as shown in Figure B 3.

Note: The purpose of this layer is to tack the Confortex fabric over the silicone layers. Apply very thin layer of silicone over the entire part and tack the Confortex fabric immediately while the silicone is wet. Overlaps in fabric may be necessary depending on the complexity of the part.

Figure B 3: Laying up the Confortex Reinforcement fabric

10. Once the third layer of silicone (silicone adhesive layer from step 9) is cured, apply a thick layer of silicone over the Confortex fabric and fully impregnate it with silicone. Allow for this layer to cure (approximately 1 hour).


11. Apply fourth and fifth layer of silicone over the bag surface, as before, to build up the thickness.

Note: The quality of the bag depends on how uniformly the silicone is brushed each time.

12. Apply another layer of Confortex fabric over the resin inlet and outlet lines and repeat Steps 9 to 11. This step is to locally reinforce the area of inlet and outlet lines to avoid the resin channels collapsing under vacuum pressure.

13. Allow the bag to cure for 24 hours before using it.

Mouldless CCBM Bae Manufacturing: Test Section and Fuselaee

1. Prepare the tool/mould : This process begins with the foam core assembly laid-up with reinforcement.

a. Wrap the entire assembly with a layer of peel ply. b. Wrap the entire assembly with a layer of PVC film (use Teflease tape as necessary

to create a tight wrap around the part). c. Install the resin inlets and outlets onto the setup using double-sided tape. d. Install the resin distribution channels using double-sided tape.

2. Follow the sequence of silicone and Confortex fabric application explained in CCBM bag manufacturing Trial #2.

Note: On large surfaces such as the fuselage walls, apply the silicone over a 2 sq. ft. region at a time. Once the region is fully covered with a layer of silicone, move on to the adjacent region. This is done to apply the silicone while it is wet.


Appendix C: Process Value Analysis

Assumptions: * Only the major costs associated with difference in the processing techniques were accounted

in this cost analysis. * Cost of standard laboratory consumables (wax, wipes, paint brushes, squeegees, etc.) were

ignored. * For comparison purposes the labour hours were converted to monetary values at 20$/hr

labour rate. * All prices were in $ CAD and were transferred directly from supplier's or manufacturer's

quotations.

* The prices do not include taxes or shipping/handling costs.

Number of parts Number of Plies

Total Area of Infusion (ft ) Total Resin Weight (lbs)

Seal Perimeter (ft)

Labour Rate ($/hr)

1 2

30 4

10

20

Table C 1: Cost factors for VARTM

Conventional VARTM (Disposable Bagging)

Item

Fabric (including 20% wastage) Resin

Distribution media (Resinflow 60)

Vacuum bag (Stretchlon 200) Peel ply (Econolease) Sealant tape (AT-200Y)

Total Cost Labour hours/part for bagging (hours) Total labour hours for bagging (hours)

Unit cost ($)

21.97/ly 350/ 52 lbs kit

100 ly/$394.30/roll/

100ly:$218.96/ea lOOly: $427.05/roll

4.195$/roll

2.5 hrx3 people

Cost ($)/ unit area (ft2) or length (ft) or weight

(lbs)

$1.83 $6.73

$0.26

$0.15 $0.28 $4.20

Total cost

$131.82 $26.92

$8.58

$12.75 $9.24 $8.39

$197.70

7.50

7.50


Table C 2: Cost factors for CCBM I

CCBM (Silicone seal) 1

Item


Silicone seal

One time part buildup cost Liquid silicone Distribution media (Resinflow 60) Peel ply (Econolease) Total cost

Labour (hours)

Unit cost ($)

21.97/ ly

350/ 52 lbs kit

9$/ft

33$/cartridge 100 ly/$394.30/roll/ lOOly: $427.05/roll

2 full days


(lbs)

$1.83 $6.73

$9.00

$33.25 $0.26 $0.28

16hrs

One time initial cost

Additional cost per part

Total cost

$131.82 $26.92

$90.00

$120.00 $332.50

$8.58 $9.24

$719.06

$542.50

$176.56

Table C 3: Cost factors for CCBM II

CCBM (Sealant tape) II

Item


One time part buildup cost

Liquid silicone Distribution media (Resinflow 60) Peel ply (Econolease) Sealant tape Total cost

Labour (hours)

Total labour (hours)

Unit cost ($)

21.97/ ly

350/52 lbs kit

33$/cartridge 100 ly/$394.30/roll/ lOOly: $427.05/roll

4.20$/roll

2- full days


(lbs)

$1.83 $6.73

$33.00 $0.26 $0.28 $4.20

15hrs



Total cost

$131.82 $26.92

$100.00

$330.00 $8.58 $9.24 $1.68

$608.24

15.50

$430.00

$178.24


Table C 4: Cost factors for CCBM III

CCBM (Integrated with distribution media) III

Item


One time part buildup cost

Liquid silicone Sealant tape Total cost Labour (hours) Total labour (hours)

Unit cost ($)

21.97/ ly 350/ 52 lbs kit

33$/cartridge 4.20$ / roll

2- full days


(lbs) $1.83 $6.73

$33.00 $4.20

15hrs



Total cost

$131.82 $26.92

$120.00

$330.00 $1.68

$610.42

15.50

$450.00

$160.42

Table C 5: Total cost calculations

Conventional bagging

Cost

$197.70

$395.40

$593.10 $790.80 $988.50

$1,186.20 $1,383.90

$1,581.60

$1,779.30

$1,977.00 $2,174.70 $2,372.40 $2,570.10 $2,767.80 $2,965.50

$3,163.20 $3,360.90 $3,558.60 $3,756.30

$3,954.00

Labour hours

7.50

15.00

22.50 30.00 37.50 45.00 52.50

60.00

67.50

75.00 82.50 90.00 97.50 105.00 112.50

120.00 127.50 135.00 142.50

150.00

CCBMI

Cost

$719.06

$895.62

$1,072.18 $1,248.74 $1,425.30 $1,601.86 $1,778.42

$1,954.98

$2,131.54

$2,308.10 $2,484.66 $2,661.22 $2,837.78 $3,014.34 $3,190.90

$3,367.46 $3,544.02 $3,720.58 $3,897.14

$4,073.70

Labour hours

16.50

17.00

17.50 18.00 18.50 19.00 19.50

20.00

20.50

21.00 21.50 22.00 22.50 23.00 23.50

24.00 24.50 25.00 25.50

26.00

CCBM II

Cost

$608.24

$786.48

$964.72 $1,142.96 $1,321.20 $1,499.44 $1,677.68

$1,855.92

$2,034.16

$2,212.40 $2,390.64

$2,568.88 $2,747.12 $2,925.36 $3,103.60

$3,281.84 $3,460.08 $3,638.32 $3,816.56

$3,994.80

Labour hours

15.50

16.00

16.50 17.00 17.50 18.00 18.50

19.00

19.50

20.00 20.50 21.00 21.50 22.00 22.50

23.00 23.50 24.00 24.50

25.00

CCBM III

Cost

$610.42

$770.84

$931.26 $1,091.68 $1,252.10 $1,412.52 $1,572.94

$1,733.36

$1,893.78

$2,054.20 $2,214.62 $2,375.04 $2,535.46 $2,695.88 $2,856.30

$3,016.72 $3,177.14 $3,337.56 $3,497.98

$3,658.40

Labour hours

16.00

16.40

16.80 17.20 17.60 18.00 18.40

18.80

19.20

19.60 20.00 20.40 20.80 21.20 21.60

22.00 22.40 22.80 23.20

23.60

Part count

1

2

3 4 5 6 7

8

9

10 11 12 13 14 15

16 17 18 19

20


Appendix D: Core Materials and Inserts

Figure D 1 to Figure D 6 show the compression, shear and tensile properties of various foam materials utilized in VARTM applications. Their costs are compared in Figure D 7.

600

500

-CELFORT® 300

-KLEGECELLR

-CORECELLT

-AIREX R 63

ROHACELL RIST

AIRCELLT

-DIVINYCELLH

-CORECELLA

-AIREX C 70

AIREXT90

-ROHACELL A

-ELFO

4 5 Density (lbs/ft3)

Figure D1: Compression strengths of foam cores


35000

30000

«o 25000 D.

I I

-CELFORT® 300

-KLEGECELLR

- CORECELL T

- AIREX R 63

AJRCELLT

-DIVINYCELLH

-CORECELLA

-AIREX C 70

AIREX T 90

-ELFQAM

3 4 5 6 Density (lbs/ft3)

Figure D 2: Compression moduli of foam cores

Figure D 3: Shear strengths of foam cores


10000

8000

-DIVINYCELLH

-CORECELLA

-AIREX C 70

AIREX T 90

3B

-KLEGECELLR

-CORECELLT

-AIREX R 63

ROHACELL RIST

UT_

3 4 , 5 Density (lbs/ft3)

Figure D 4: Shear moduli of foam cores

800

700

600

i

-DIVINYCELLH

-CORECELLA

-AIREX C 70

ROHACELL RIST

-KLEGECELLR

-CORECELLT

-AIREX R 63

-ROHACELLA

3 4 5 Density (lbs/ft3)

Figure D 5: Tensile strengths of foam cores


30000

25000

'5 20000 Q.

0)

3 "g 15000

(0

« 10000

5000


Figure D 6: Tensile moduli of foam cores

16

14

12 -

10 -

o 'C Q.

6 -

4

2

0 -

1"+! PU

T 1

PET PVC-Cross Linked

SAN PMI

* Prices are for 8 ft x 4 ft, 0.5 in thick sheets, averaged over density range of 3 lbs/ft to 8 lbs/ft * All values are based retail market prices as of Aug08, obtained either from the manufacturers or the

local distributors

Figure D 7: Average prices of various foam cores


Figure D 8 to Figure D 13 show the properties of balsa cores in comparision to the structural

PMI (Rohacell) foams. The costs of balsa and foam cores are compared in Figure D 14.

4000

3500

s=-3000 a

3)2500 c a> i^

"> 2000 c o » 2 1500 a E O 1000

500

I ^ ^ ROHACELL RIST

^-DIVINYCELLBULSA

-•-BALTEKBULSA

J '/ *

/ /

/ / y

/ *

/


14 16

Figure D 8: Compression strengths of balsa cores with reference to PMI foam core


•a o

in 0) Q. E o o

1000000

900000

800000

700000

600000

500000

400000

300000

200000

100000

0

! I ! -•-CORECELLS

- o - DIVINYCELL BULSA

-•-BALTEKBULSA

•HBB

r?

y / /

/ •

/

6 8 1

Density (lbs/ft3)

0 12 14 16

Figure D 9: Compression moduli of balsa cores with reference to SAN foam core

800

700

600

(0

3 500

O)

c

2 (A w CO

a £

400

300

200

100

-ROHACELL RIST

-DIVINYCELL BULSA

-BALTEKBULSA

6 8 10 12 14 16 Density (lbs/ft3)

Figure D10: Shear strengths of balsa cores with reference to PMI foam core


<n O.

50000

45000

40000

N35000

'• 30000 w 3 •o 25000 o 2 jj 20000 4) £

w 15000

10000

5000

0

- o - ROHACELL RIST

-^-DIVINYCELL BULSA

-^BALTJEKBULSft

^

O^ ^

0 ^

^

/ / ^

/

6 8 1( Density (lbs/ft3)

12 14 16

Figure D 11: Shear moduli of balsa cores with reference to PMI foam core

4000

3500

3000

0)

S2500

O)

® 2000 CO

« 35 1500 c

1000

500

- ° - ROHACELL RIST

- o - DIVINYCELL BULSA

-•-BALTEKBULSA

~*

/


12 14 16

Figure D12: Tensile strengths of balsa cores with reference to PMI foam core


900000

800000

700000

•</> 600000 a w • | 500000 •D O S 400000 0)

'55 S 300000

200000

100000

i

-a-ROHACELL RIST

D

^


12 14 16

Figure D 13: Tensile moduli of balsa cores with reference to PMI foam core

O

16

14

12

10

8

6

4

2

0

6

- f c

Foam Cores

d ^

-0F

Balsa Core

Figure D14: Cost of balsa core compared to foam cores


Material Properties of Core Materials:

1. Divinycell: Diab Corporation http://www.diabgroup.eom/europe/products/e prods l.html

2. Alcan Airex: Alcan Composites http://www.corematerials.3acomposites.com/america.html ?&L=0

3. Nida Core PET, PU, Foam Filled Honey Comb: Nida Core Corporation http://www.nida-core.com/english/index.htm

4. Aircell Polyester Foam Cores-Polyumac Inc. http://www.polyumac.com/Aircellmp.htm 5. Elfoam: Elliott Company- http://www.elliottfoam.com/tech.html 6. Corecell: SP Systems- http://www.marineware.com/ccp 2.asp 7. Rohacell: Evonik Industries /Degussa

http://www.rohacell.com/en/performanceplastics8344.html 8. Last-A-Foam: General Plastics Inc.

http://www.generalplastics.com/products/product detail.php?pid=19

Material Properties of Inserts:

1. FRP inserts, PEEK, Glass Filled Peek: Mc-Master Carr, http://www.mcmaster.eom/#


http://www.diabgroup.eom/europe/products/e

http://www.corematerials.3acomposites.com/america.html

http://www.nidacore.com/english/index.htm

http://www.nidacore.com/english/index.htm

http://www.polyumac.com/Aircellmp.htm

http://www.elliottfoam.com/tech.html

http://www.marineware.com/ccp

http://www.rohacell.com/en/performanceplastics8344.html

http://www.generalplastics.com/products/product

http://www.mcmaster.eom/%23

Appendix E: FEA Results and Weight Estimates

FEA Results:

Table E 1: FEA Results: In-Flight Analysis Step

Analysis Step

In-Flight

In-Flight

In-Flight

In-Flight

Location

Carry-through Spar

Shear Pin

Rear Bulkhead

Front Bulkhead

Max. Von Mises Stress

(psi) 60,000

40, 000

69,000

20,000

Max. In-Plane Shear Stress,

S12 (psi)

13,000

6,500

9,500

4,000

Max In-Plane Strain. (IE)

0.010

0.004

0.008

0.003

Table E 2: FEA Results: Landing analysis step

Analysis Step

Landing

Landing

Landing

Landing

Landing

Location

Carry-through Spar

Shear Pin

Rear Bulkhead

Front Bulkhead

Landing Gear Attachment

Max. Von Mises Stress

(psi)

34, 000

10, 000

30, 000

28, 000

52,000

Max. In-Plane Shear Stress,

S« (psi)

3,500

2,500

4,000

7,000

13,500

Max In-Plane Strain. (13)

0.007

0.002

0.003

0.005

0.011

Weight Estimates:

Table E 3: Weights of fuselage panels predicted by the FEA

*(lbs)*

Skin Core

Base plate

0.79 0.16

Side walls

1.54 0.67

Fairings

0.29 0.73

Front bulkhead

0.29 0.17

Rear bulkhead

0.44 0.31

Bushings

1.71 N/A

Landing gear

bracket 1.99 0.42

Nose cone

bridge 0.15 0.07

Rear side walls 0.92 0.46

Bay separator

panel 0.17 0.05

Appendix E: FEA Results and Weight Estimates 157

Appendix F: Microscopic Image Analysis

VARTM Specimens:

Good = Partially open cells at the skin-core interface are filled with resin.

Bad = Voids at the skin core interface.

Figure F1: Sample cross-sections of VARTM specimens


CCBM Specimens:

Figure F 2: Images of the CCBM specimens at random locations


Appendix G: Fuselage Profiling

The outer mould line (OML) (Figure G 1) of the new fuselage was measured around all of the

edges with particular attention to the regions that interface with the access panels. The entire

process and results are discussed in the following few sections.

\- 20 81

Figure G 1: Overall dimensions of the fuselage assembly


Procedure:

The profiling of the fuselage was carried out in a process involving several steps. These steps

are outlined below.

1. The Pro/E model of the assembled fuselage (without the carbon fibre skin) was examined and all of the points at which measurements would be taken were determined.

2. The dimensions of the Pro/E model at each of these points were recorded. Experimental skin thickness was added to the Pro/E dimensions to obtain the ideal dimensions listed in Table G 1 through Table G 6.

3. The fuselage was marked with masking tape at each of the points where measurements would be taken, as can be seen in Figure G 2 and Figure G 3.

4. All of the measurements of the fuselage were taken with a shop ruler to a precision of at least ±1/16".

5. All of the data was inserted into a MS Excel™ template to obtain the deviations found in Table G 1 through Table G 6 and to determine the tolerances for each set of results.

6. The results were used to sketch the actual profiles of the fuselage on top of the ideal profiles as shown in Figure G 4 and Figure G 5.


> T3 T3 <T> Q_ x '

c (/> ro_

era

3 era

NOTES ' Measurements A through AP are taken normal to the planes they are shown on

' Points A and X are 1 5" frorr. the rear of the rear buikheac • Points A-D and W-T are measured in 5" increments from the rear oT the rear bulkhead

' Points G-J and O-R ere measured in 5" increments from the front bulkhead along ther respective edges

' Points AE and AK align with point I * Measurements AS and AT are lire put equvalenls tg they are shown on measurements AW and AX

* _ / W_7 v / U__/ T_j !_§ R_/ Q / CD M

»

Figure G 3: Marking the fuselage with masking tape to identify the points shown in Figure G 2

Results:

Table G 1 through Table G 6 show the measured values at the locations illustrated in Figure G 2. The results are split into separate tables according to which part of the fuselage they pertain to. The deviation columns are colour coded by the magnitude of the deviation with the darker red showing the greatest deviation and the brighter green being showing the smallest.

Note that the fuselage widths along the top and the bottom surfaces have the greatest deviations. This is of most interest for the Structures Fuselage Team, since these are the areas that pertain to the attachment of the access panels. Figure G 4 and Figure G 5 show precisely where these measurements were taken and also the show spline curves (in red) connecting where the measurements were taken, in order to give the complete shape profile of the actual fuselage. Note that in creating these spline curves it was assumed that the deviation is symmetric about the centreline of the fuselage. This assumption is not perfect, but is adequate for the purpose of demonstrating the deviation.


Table G 1: Width results across the top of the fuselage

Measurement ID

A B C D E F G H 1 J K

Measurement ID

0 p Q R S T U V

w X

Measurement [in.]

(±1/16)

13 8/32 13 0/32 12 30/32 13 0/32 13 4/32 13 12/32 13 2/32 12 20/32 12 5/32 11 21/32 117/32

Table G 2: Width results

Measurement [in.]

(+1/16) 11 8/32 11 20/32 12 8/32 12 30/32 13 14/32 13 10/32 13 14/32 13 10/32 13 14/32 13 13/32

Ideal [in.] A-F (±0.010) G-K (±0.006)

13.34 13.34 13.34 13.34 13.34 13.34 13.22 12.75 12.28 11.81 11.34

Deviation [in.] A-F (±0.073) G-K (±0.069)

-0.09 -0.34 -0.40 -0.34 -0.21 0.04 -0.16 -0.12 -0.12 -0.16 -0.12

across the bottom of the fuselage

Ideal [in.] (±0.006)

11.33 11.88 12.36 12.83

13.34+0.010 13.34±0.010 13.38±0.018 13.38±0.018 13.48±0.034 13.48±0.034

Deviation [in.] (±0.069)

-0.08 -0.26 -0.11 0.11

0.10±0.073 -0.02+0.073 0.05+0.081 -0.07+0.081 -0.04+0.097 -0.07+0.097


22.19

1.5

K 11-7/32 J

11-21/32

I 12-5/32

H 12-5/8

G 13-1/16

F 13-3/8

E 13-1/8

D 43-

C 12-15/16

B 13

A 13-1/4

5.0

- M 0 . 0

15.0

20.0

5

/

20.0

15.0

T 5.0

J_

10.0

Figure G 4: Outline of top of fuselage showing the actual (in red) and the ideal profiles and the locations of the measurement points shown in Table G 1. The dimensions shown are the measured dimensions. The positions of the measurements are

shown relative to the fuselage Outer Mould Line (OML).


22.19

Figure G 5: Outline of bottom of fuselage showing the actual (in red) and the ideal profiles and the locations of the measurement points shown in Table G 2. The dimensions shown are the measured dimensions. The positions of the

measurements are shown relative to the fuselage Outer Mould Line (OML).


Table G 3: Front bulkhead height and width dimensions

Measurement ID Measurement [in.]

(+1/16)

Ideal [in.] (±0.01)

10.93 10.93 10.93 14.22 15.10 14.22


-0.05 -0.05 -0.05 -0.06 -0.03 -0.06

L M N

AG AH Al

10 28/32 10 28/32 10 28/32 14 5/32 15 2/32 14 5/32

Table G 4: Rear bulkhead height and width dimensions


(±1/16)

Ideal [in.] Deviation [in.]

Y Z

AA AO AP

13 12/32 13 10/32 13 10/32 15 16/32 15 16/32

13.50+0.034 13.34+0.014 13.34+0.014 15.52+0.022 15.52+0.022

-0.13+0.097 -0.05+0.077 -0.05+0.077 -0.02+0.085 -0.02+0.085

Table G 5: Front bulkhead tab dimensions

Measurement ID Measurement [in.] (±0.001)

Ideal [in.] (±0.01)


AY

AZ BA BB BC

4.012 0.907

0.386 0.374 0.512

4.035

0.895 0.355

0.355 0.500+0.00

-0.044 -0.009

0.010 -0.002 0.012

Table G 6: Fuselage length results


(± 1/16)

Ideal [in.] Deviation [in.]

AQ AR AS AT AU AV AW AX

22 5/16 20 13/16 22 5/16

20 13/16 22 4/16

20 13/16 22 5/16

20 13/16

22.23+0.007 20.97+0.005 22.29+0.017 20.82+0.005 22.24+0.007

20.97+0.005 22.29+0.017 20.82+0.005

0.07+0.070 -0.16+0.068 0.02+0.080 -0.01+0.068 0.01+0.070 -0.16+0.068 0.02+0.080 0.07+0.068


Appendix H: Fibre Volume Fraction Calculation

Weight of resin infused (lbs)

Likewise,

Weight of fabric on the fuselage (lbs)=

Pi 300 fibres

V. skin

2.916 lbs_

i3

0.0425 ft3

68.64 —j, where pn ft

M.

K,

4.884

110.6 Ibs^

ft' 0.0442 ft3

Using, Vresm+Vfabnc = Vskin, (zero void content is assumed, and weave effects of the fabric is

ignored)

Vf ~ 51%.

Appendix H: Fibre Volume Fraction Calculation 168